 TensorFlow.orgで表示 |
 Google Colab で実行 |
 GitHub でソースを表示{ |
概要
このチュートリアルでは、Neural Structured Learning（NSL）フレームワークを用いた画像分類のための敵対的学習（Goodfellow et al.、2014）の使用について説明します。
敵対的学習の中核となる考え方は、有機的な学習データに加え、敵対的なデータ（敵対的サンプルと呼ばれる）を用いてモデルをトレーニングすることです。これらの敵対的サンプルは人間の目には元のデータと同じように見えますが、摂動がモデルを混乱させ、誤った予測や分類が行われる原因となります。敵対的サンプルはモデルに誤った予測や分類を行わせ、誤認識させるように意図的に構成されています。このような例を用いてトレーニングを行うことによって、モデルは予測を行う際に敵対的摂動に対してロバストになるよう学習します。
このチュートリアルでは、Neural Structured Learning のフレームワークを使用し、ロバストなモデルを得るために敵対的学習を適用する手順を以下に説明します。
- 基本モデルとしてニューラルネットワークを作成します。このチュートリアルでは、
tf.kerasFunctional API で基本モデルを作成します。この手順は、tf.kerasSequential API および Subclassing API で作成されたモデルと互換性があります。TensorFlow における Keras モデルの詳細については、こちらのドキュメントをご覧ください。 - NSL フレームワークが提供するラッパークラス
AdversarialRegularizationで基本モデルをラップして、新しいtf.keras.Modelインスタンスを作成します。この新しいモデルには、トレーニング目的の正則化項として敵対的損失が含まれます。 - トレーニングデータの例を特徴ディクショナリに変換します。
- 新しいモデルをトレーニングして評価します。
セットアップ
Neural Structured Learning パッケージをインストールします。
pip install --quiet neural-structured-learningライブラリをインポートします。neural_structured_learning を nsl と略します。
import matplotlib.pyplot as plt
import neural_structured_learning as nsl
import numpy as np
import tensorflow as tf
import tensorflow_datasets as tfds
ハイパーパラメータ
モデルのトレーニングと評価のための（HParams オブジェクト内の）ハイパーパラメータを集めて説明します。
入力/出力：
input_shape：入力テンソルの形状。各画像は 28×28 ピクセルで 1 チャンネルです。num_classes：[0-9] に対応する数字 10 個分があり、合計 10 クラスです。
モデルアーキテクチャ：
conv_filters：各畳み込みレイヤーのフィルタ数を指定する数値のリスト。kernel_size：すべての畳み込みレイヤーで共有する 2 次元の畳み込みウィンドウのサイズ。pool_size：各最大 Pooling レイヤーで画像をダウンスケールするための係数。num_fc_units：完全に接続された各レイヤーの単位（すなわち幅）の数。
トレーニングと評価：
batch_size：トレーニングや評価に使用するバッチサイズ。epochs：トレーニングのエポック数。
敵対的学習：
adv_multiplier：ラベル付けされた損失に対して相対的な、トレーニング目的内の敵対的損失の重み。adv_step_size：敵対的摂動の大きさ。adv_grad_norm：敵対的摂動の大きさを測るノルム。
class HParams(object):
def __init__(self):
self.input_shape = [28, 28, 1]
self.num_classes = 10
self.conv_filters = [32, 64, 64]
self.kernel_size = (3, 3)
self.pool_size = (2, 2)
self.num_fc_units = [64]
self.batch_size = 32
self.epochs = 5
self.adv_multiplier = 0.2
self.adv_step_size = 0.2
self.adv_grad_norm = 'infinity'
HPARAMS = HParams()
MNIST データセット
MNIST データセットには手書きの数字（ '0' から '9' まで）のグレースケール画像が含まれています。各画像は低解像度（28×28 ピクセル）で数字 1 文字を示しています。このタスクでは、画像を数字ごとに 1 つずつ、10 のカテゴリに分類します。
ここで TensorFlow Datasets から MNIST データセットを読み込みます。これはデータのダウンロードと tf.data.Dataset の構築処理をします。読み込んだデータセットには 2 つのサブセットがあります。
- 6 万個の例を含む
train - 1 万個の例を含む
test
両方のサブセットに含まれる例は、以下の 2 つのキーを持つ特徴ディクショナリに格納されています。
image：0 から 255 までのピクセル値の配列。label：0 から 9までの真の正解ラベル。
datasets = tfds.load('mnist')
train_dataset = datasets['train']
test_dataset = datasets['test']
IMAGE_INPUT_NAME = 'image'
LABEL_INPUT_NAME = 'label'
モデルを数値的に安定させるには、normalize 関数にデータセットをマッピングし、ピクセル値を [0, 1] に正規化します。トレーニングセットをシャッフルしてバッチ処理を行った後、基本モデルをトレーニングするために例を特徴タプル (image, label) に変換します。また、後で使用できるようにタプルをディクショナリに変換する関数も用意しています。
def normalize(features):
features[IMAGE_INPUT_NAME] = tf.cast(
features[IMAGE_INPUT_NAME], dtype=tf.float32) / 255.0
return features
def convert_to_tuples(features):
return features[IMAGE_INPUT_NAME], features[LABEL_INPUT_NAME]
def convert_to_dictionaries(image, label):
return {IMAGE_INPUT_NAME: image, LABEL_INPUT_NAME: label}
train_dataset = train_dataset.map(normalize).shuffle(10000).batch(HPARAMS.batch_size).map(convert_to_tuples)
test_dataset = test_dataset.map(normalize).batch(HPARAMS.batch_size).map(convert_to_tuples)
基本モデル
基本モデルは、3 つの畳み込みレイヤーと 2 つの完全に接続されたレイヤー（HPARAMS で定義されている）から成るニューラルネットワークです。ここでは Keras の Functional API を使用して定義しています。他の API やモデルアーキテクチャを自由に試してみてください。
def build_base_model(hparams):
"""Builds a model according to the architecture defined in `hparams`."""
inputs = tf.keras.Input(
shape=hparams.input_shape, dtype=tf.float32, name=IMAGE_INPUT_NAME)
x = inputs
for i, num_filters in enumerate(hparams.conv_filters):
x = tf.keras.layers.Conv2D(
num_filters, hparams.kernel_size, activation='relu')(
x)
if i < len(hparams.conv_filters) - 1:
# max pooling between convolutional layers
x = tf.keras.layers.MaxPooling2D(hparams.pool_size)(x)
x = tf.keras.layers.Flatten()(x)
for num_units in hparams.num_fc_units:
x = tf.keras.layers.Dense(num_units, activation='relu')(x)
pred = tf.keras.layers.Dense(hparams.num_classes, activation='softmax')(x)
model = tf.keras.Model(inputs=inputs, outputs=pred)
return model
base_model = build_base_model(HPARAMS)
base_model.summary()
Model: "model" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= image (InputLayer) [(None, 28, 28, 1)] 0 _________________________________________________________________ conv2d (Conv2D) (None, 26, 26, 32) 320 _________________________________________________________________ max_pooling2d (MaxPooling2D) (None, 13, 13, 32) 0 _________________________________________________________________ conv2d_1 (Conv2D) (None, 11, 11, 64) 18496 _________________________________________________________________ max_pooling2d_1 (MaxPooling2 (None, 5, 5, 64) 0 _________________________________________________________________ conv2d_2 (Conv2D) (None, 3, 3, 64) 36928 _________________________________________________________________ flatten (Flatten) (None, 576) 0 _________________________________________________________________ dense (Dense) (None, 64) 36928 _________________________________________________________________ dense_1 (Dense) (None, 10) 650 ================================================================= Total params: 93,322 Trainable params: 93,322 Non-trainable params: 0 _________________________________________________________________
次に、基本モデルのトレーニングと評価を行います。
base_model.compile(optimizer='adam', loss='sparse_categorical_crossentropy',
metrics=['acc'])
base_model.fit(train_dataset, epochs=HPARAMS.epochs)
Epoch 1/5 1/1875 [..............................] - ETA: 1:40:15 - loss: 2.3070 - acc: 0.1250 9/1875 [..............................] - ETA: 12s - loss: 2.2814 - acc: 0.2203 18/1875 [..............................] - ETA: 11s - loss: 2.2622 - acc: 0.2192 27/1875 [..............................] - ETA: 11s - loss: 2.2154 - acc: 0.2483 36/1875 [..............................] - ETA: 11s - loss: 2.1397 - acc: 0.2847 45/1875 [..............................] - ETA: 10s - loss: 2.0564 - acc: 0.3192 54/1875 [..............................] - ETA: 10s - loss: 1.9715 - acc: 0.3512 64/1875 [>.............................] - ETA: 10s - loss: 1.8808 - acc: 0.3845 73/1875 [>.............................] - ETA: 10s - loss: 1.8052 - acc: 0.4117 82/1875 [>.............................] - ETA: 10s - loss: 1.7353 - acc: 0.4364 91/1875 [>.............................] - ETA: 10s - loss: 1.6705 - acc: 0.4588 100/1875 [>.............................] - ETA: 10s - loss: 1.6113 - acc: 0.4791 110/1875 [>.............................] - ETA: 10s - loss: 1.5520 - acc: 0.4992 119/1875 [>.............................] - ETA: 10s - loss: 1.5032 - acc: 0.5156 128/1875 [=>............................] - ETA: 10s - loss: 1.4583 - acc: 0.5307 138/1875 [=>............................] - ETA: 9s - loss: 1.4123 - acc: 0.5461 148/1875 [=>............................] - ETA: 9s - loss: 1.3699 - acc: 0.5603 157/1875 [=>............................] - ETA: 9s - loss: 1.3344 - acc: 0.5721 166/1875 [=>............................] - ETA: 9s - loss: 1.3015 - acc: 0.5829 175/1875 [=>............................] - ETA: 9s - loss: 1.2708 - acc: 0.5931 184/1875 [=>............................] - ETA: 9s - loss: 1.2420 - acc: 0.6026 193/1875 [==>...........................] - ETA: 9s - loss: 1.2147 - acc: 0.6116 202/1875 [==>...........................] - ETA: 9s - loss: 1.1891 - acc: 0.6200 212/1875 [==>...........................] - ETA: 9s - loss: 1.1623 - acc: 0.6288 221/1875 [==>...........................] - ETA: 9s - loss: 1.1395 - acc: 0.6362 230/1875 [==>...........................] - ETA: 9s - loss: 1.1178 - acc: 0.6433 240/1875 [==>...........................] - ETA: 9s - loss: 1.0951 - acc: 0.6507 250/1875 [===>..........................] - ETA: 9s - loss: 1.0736 - acc: 0.6577 259/1875 [===>..........................] - ETA: 9s - loss: 1.0551 - acc: 0.6637 268/1875 [===>..........................] - ETA: 9s - loss: 1.0374 - acc: 0.6694 278/1875 [===>..........................] - ETA: 9s - loss: 1.0187 - acc: 0.6754 287/1875 [===>..........................] - ETA: 9s - loss: 1.0027 - acc: 0.6806 296/1875 [===>..........................] - ETA: 8s - loss: 0.9873 - acc: 0.6856 305/1875 [===>..........................] - ETA: 8s - loss: 0.9726 - acc: 0.6904 314/1875 [====>.........................] - ETA: 8s - loss: 0.9584 - acc: 0.6950 323/1875 [====>.........................] - ETA: 8s - loss: 0.9447 - acc: 0.6994 332/1875 [====>.........................] - ETA: 8s - loss: 0.9315 - acc: 0.7036 341/1875 [====>.........................] - ETA: 8s - loss: 0.9188 - acc: 0.7077 350/1875 [====>.........................] - ETA: 8s - loss: 0.9065 - acc: 0.7117 360/1875 [====>.........................] - ETA: 8s - loss: 0.8934 - acc: 0.7159 370/1875 [====>.........................] - ETA: 8s - loss: 0.8807 - acc: 0.7199 379/1875 [=====>........................] - ETA: 8s - loss: 0.8697 - acc: 0.7235 389/1875 [=====>........................] - ETA: 8s - loss: 0.8580 - acc: 0.7272 399/1875 [=====>........................] - ETA: 8s - loss: 0.8467 - acc: 0.7309 409/1875 [=====>........................] - ETA: 8s - loss: 0.8357 - acc: 0.7344 418/1875 [=====>........................] - ETA: 8s - loss: 0.8261 - acc: 0.7374 427/1875 [=====>........................] - ETA: 8s - loss: 0.8169 - acc: 0.7404 437/1875 [=====>........................] - ETA: 8s - loss: 0.8069 - acc: 0.7436 447/1875 [======>.......................] - ETA: 8s - loss: 0.7972 - acc: 0.7467 456/1875 [======>.......................] - ETA: 8s - loss: 0.7887 - acc: 0.7494 466/1875 [======>.......................] - ETA: 8s - loss: 0.7796 - acc: 0.7523 476/1875 [======>.......................] - ETA: 7s - loss: 0.7708 - acc: 0.7552 485/1875 [======>.......................] - ETA: 7s - loss: 0.7630 - acc: 0.7576 494/1875 [======>.......................] - ETA: 7s - loss: 0.7555 - acc: 0.7600 503/1875 [=======>......................] - ETA: 7s - loss: 0.7482 - acc: 0.7624 512/1875 [=======>......................] - ETA: 7s - loss: 0.7411 - acc: 0.7647 521/1875 [=======>......................] - ETA: 7s - loss: 0.7341 - acc: 0.7669 530/1875 [=======>......................] - ETA: 7s - loss: 0.7274 - acc: 0.7691 540/1875 [=======>......................] - ETA: 7s - loss: 0.7200 - acc: 0.7714 550/1875 [=======>......................] - ETA: 7s - loss: 0.7129 - acc: 0.7737 560/1875 [=======>......................] - ETA: 7s - loss: 0.7059 - acc: 0.7759 569/1875 [========>.....................] - ETA: 7s - loss: 0.6998 - acc: 0.7779 578/1875 [========>.....................] - ETA: 7s - loss: 0.6938 - acc: 0.7798 588/1875 [========>.....................] - ETA: 7s - loss: 0.6873 - acc: 0.7819 597/1875 [========>.....................] - ETA: 7s - loss: 0.6816 - acc: 0.7837 606/1875 [========>.....................] - ETA: 7s - loss: 0.6760 - acc: 0.7855 615/1875 [========>.....................] - ETA: 7s - loss: 0.6706 - acc: 0.7873 625/1875 [=========>....................] - ETA: 7s - loss: 0.6647 - acc: 0.7892 634/1875 [=========>....................] - ETA: 7s - loss: 0.6595 - acc: 0.7908 643/1875 [=========>....................] - ETA: 6s - loss: 0.6543 - acc: 0.7925 652/1875 [=========>....................] - ETA: 6s - loss: 0.6493 - acc: 0.7941 662/1875 [=========>....................] - ETA: 6s - loss: 0.6439 - acc: 0.7958 672/1875 [=========>....................] - ETA: 6s - loss: 0.6386 - acc: 0.7975 681/1875 [=========>....................] - ETA: 6s - loss: 0.6339 - acc: 0.7990 691/1875 [==========>...................] - ETA: 6s - loss: 0.6288 - acc: 0.8007 700/1875 [==========>...................] - ETA: 6s - loss: 0.6242 - acc: 0.8021 710/1875 [==========>...................] - ETA: 6s - loss: 0.6193 - acc: 0.8037 719/1875 [==========>...................] - ETA: 6s - loss: 0.6150 - acc: 0.8051 728/1875 [==========>...................] - ETA: 6s - loss: 0.6108 - acc: 0.8064 737/1875 [==========>...................] - ETA: 6s - loss: 0.6066 - acc: 0.8077 746/1875 [==========>...................] - ETA: 6s - loss: 0.6025 - acc: 0.8091 755/1875 [===========>..................] - ETA: 6s - loss: 0.5985 - acc: 0.8103 765/1875 [===========>..................] - ETA: 6s - loss: 0.5941 - acc: 0.8118 774/1875 [===========>..................] - ETA: 6s - loss: 0.5902 - acc: 0.8130 783/1875 [===========>..................] - ETA: 6s - loss: 0.5864 - acc: 0.8142 792/1875 [===========>..................] - ETA: 6s - loss: 0.5826 - acc: 0.8154 801/1875 [===========>..................] - ETA: 6s - loss: 0.5789 - acc: 0.8166 810/1875 [===========>..................] - ETA: 6s - loss: 0.5753 - acc: 0.8178 820/1875 [============>.................] - ETA: 5s - loss: 0.5713 - acc: 0.8190 830/1875 [============>.................] - ETA: 5s - loss: 0.5675 - acc: 0.8203 841/1875 [============>.................] - ETA: 5s - loss: 0.5633 - acc: 0.8216 852/1875 [============>.................] - ETA: 5s - loss: 0.5592 - acc: 0.8229 862/1875 [============>.................] - ETA: 5s - loss: 0.5555 - acc: 0.8241 872/1875 [============>.................] - ETA: 5s - loss: 0.5519 - acc: 0.8252 882/1875 [=============>................] - ETA: 5s - loss: 0.5484 - acc: 0.8264 892/1875 [=============>................] - ETA: 5s - loss: 0.5449 - acc: 0.8275 902/1875 [=============>................] - ETA: 5s - loss: 0.5414 - acc: 0.8286 912/1875 [=============>................] - ETA: 5s - loss: 0.5381 - acc: 0.8297 922/1875 [=============>................] - ETA: 5s - loss: 0.5348 - acc: 0.8307 932/1875 [=============>................] - ETA: 5s - loss: 0.5315 - acc: 0.8318 941/1875 [==============>...............] - ETA: 5s - loss: 0.5286 - acc: 0.8327 951/1875 [==============>...............] - ETA: 5s - loss: 0.5254 - acc: 0.8337 961/1875 [==============>...............] - ETA: 5s - loss: 0.5223 - acc: 0.8347 971/1875 [==============>...............] - ETA: 5s - loss: 0.5192 - acc: 0.8357 981/1875 [==============>...............] - ETA: 5s - loss: 0.5162 - acc: 0.8366 991/1875 [==============>...............] - ETA: 4s - loss: 0.5133 - acc: 0.8376 1001/1875 [===============>..............] - ETA: 4s - loss: 0.5103 - acc: 0.8385 1011/1875 [===============>..............] - ETA: 4s - loss: 0.5075 - acc: 0.8394 1021/1875 [===============>..............] - ETA: 4s - loss: 0.5046 - acc: 0.8403 1031/1875 [===============>..............] - ETA: 4s - loss: 0.5019 - acc: 0.8412 1041/1875 [===============>..............] - ETA: 4s - loss: 0.4991 - acc: 0.8421 1051/1875 [===============>..............] - ETA: 4s - loss: 0.4964 - acc: 0.8430 1061/1875 [===============>..............] - ETA: 4s - loss: 0.4938 - acc: 0.8438 1071/1875 [================>.............] - ETA: 4s - loss: 0.4911 - acc: 0.8447 1081/1875 [================>.............] - ETA: 4s - loss: 0.4886 - acc: 0.8455 1091/1875 [================>.............] - ETA: 4s - loss: 0.4860 - acc: 0.8463 1101/1875 [================>.............] - ETA: 4s - loss: 0.4835 - acc: 0.8471 1111/1875 [================>.............] - ETA: 4s - loss: 0.4810 - acc: 0.8479 1121/1875 [================>.............] - ETA: 4s - loss: 0.4786 - acc: 0.8487 1131/1875 [=================>............] - ETA: 4s - loss: 0.4761 - acc: 0.8495 1140/1875 [=================>............] - ETA: 4s - loss: 0.4740 - acc: 0.8501 1150/1875 [=================>............] - ETA: 4s - loss: 0.4716 - acc: 0.8509 1160/1875 [=================>............] - ETA: 3s - loss: 0.4693 - acc: 0.8516 1170/1875 [=================>............] - ETA: 3s - loss: 0.4670 - acc: 0.8524 1179/1875 [=================>............] - ETA: 3s - loss: 0.4650 - acc: 0.8530 1189/1875 [==================>...........] - ETA: 3s - loss: 0.4628 - acc: 0.8537 1199/1875 [==================>...........] - ETA: 3s - loss: 0.4606 - acc: 0.8544 1208/1875 [==================>...........] - ETA: 3s - loss: 0.4586 - acc: 0.8551 1217/1875 [==================>...........] - ETA: 3s - loss: 0.4567 - acc: 0.8557 1227/1875 [==================>...........] - ETA: 3s - loss: 0.4546 - acc: 0.8564 1237/1875 [==================>...........] - ETA: 3s - loss: 0.4525 - acc: 0.8570 1247/1875 [==================>...........] - ETA: 3s - loss: 0.4504 - acc: 0.8577 1256/1875 [===================>..........] - ETA: 3s - loss: 0.4485 - acc: 0.8583 1266/1875 [===================>..........] - ETA: 3s - loss: 0.4465 - acc: 0.8589 1276/1875 [===================>..........] - ETA: 3s - loss: 0.4445 - acc: 0.8596 1286/1875 [===================>..........] - ETA: 3s - loss: 0.4425 - acc: 0.8602 1296/1875 [===================>..........] - ETA: 3s - loss: 0.4406 - acc: 0.8608 1306/1875 [===================>..........] - ETA: 3s - loss: 0.4387 - acc: 0.8615 1316/1875 [====================>.........] - ETA: 3s - loss: 0.4368 - acc: 0.8621 1326/1875 [====================>.........] - ETA: 3s - loss: 0.4349 - acc: 0.8627 1336/1875 [====================>.........] - ETA: 2s - loss: 0.4330 - acc: 0.8633 1346/1875 [====================>.........] - ETA: 2s - loss: 0.4312 - acc: 0.8639 1356/1875 [====================>.........] - ETA: 2s - loss: 0.4294 - acc: 0.8644 1366/1875 [====================>.........] - ETA: 2s - loss: 0.4276 - acc: 0.8650 1376/1875 [=====================>........] - ETA: 2s - loss: 0.4258 - acc: 0.8656 1386/1875 [=====================>........] - ETA: 2s - loss: 0.4240 - acc: 0.8661 1396/1875 [=====================>........] - ETA: 2s - loss: 0.4223 - acc: 0.8667 1406/1875 [=====================>........] - ETA: 2s - loss: 0.4206 - acc: 0.8673 1416/1875 [=====================>........] - ETA: 2s - loss: 0.4189 - acc: 0.8678 1426/1875 [=====================>........] - ETA: 2s - loss: 0.4172 - acc: 0.8683 1436/1875 [=====================>........] - ETA: 2s - loss: 0.4155 - acc: 0.8689 1445/1875 [======================>.......] - ETA: 2s - loss: 0.4140 - acc: 0.8693 1455/1875 [======================>.......] - ETA: 2s - loss: 0.4124 - acc: 0.8699 1465/1875 [======================>.......] - ETA: 2s - loss: 0.4108 - acc: 0.8704 1475/1875 [======================>.......] - ETA: 2s - loss: 0.4092 - acc: 0.8709 1484/1875 [======================>.......] - ETA: 2s - loss: 0.4078 - acc: 0.8714 1494/1875 [======================>.......] - ETA: 2s - loss: 0.4062 - acc: 0.8719 1503/1875 [=======================>......] - ETA: 2s - loss: 0.4048 - acc: 0.8723 1513/1875 [=======================>......] - ETA: 1s - loss: 0.4033 - acc: 0.8728 1523/1875 [=======================>......] - ETA: 1s - loss: 0.4018 - acc: 0.8733 1533/1875 [=======================>......] - ETA: 1s - loss: 0.4003 - acc: 0.8738 1543/1875 [=======================>......] - ETA: 1s - loss: 0.3988 - acc: 0.8742 1553/1875 [=======================>......] - ETA: 1s - loss: 0.3973 - acc: 0.8747 1563/1875 [========================>.....] - ETA: 1s - loss: 0.3958 - acc: 0.8752 1579/1875 [========================>.....] - ETA: 1s - loss: 0.3935 - acc: 0.8759 1596/1875 [========================>.....] - ETA: 1s - loss: 0.3911 - acc: 0.8767 1613/1875 [========================>.....] - ETA: 1s - loss: 0.3887 - acc: 0.8775 1629/1875 [=========================>....] - ETA: 1s - loss: 0.3865 - acc: 0.8782 1645/1875 [=========================>....] - ETA: 1s - loss: 0.3843 - acc: 0.8789 1662/1875 [=========================>....] - ETA: 1s - loss: 0.3821 - acc: 0.8796 1679/1875 [=========================>....] - ETA: 1s - loss: 0.3799 - acc: 0.8803 1695/1875 [==========================>...] - ETA: 0s - loss: 0.3778 - acc: 0.8810 1712/1875 [==========================>...] - ETA: 0s - loss: 0.3756 - acc: 0.8817 1728/1875 [==========================>...] - ETA: 0s - loss: 0.3736 - acc: 0.8823 1744/1875 [==========================>...] - ETA: 0s - loss: 0.3716 - acc: 0.8829 1760/1875 [===========================>..] - ETA: 0s - loss: 0.3697 - acc: 0.8836 1776/1875 [===========================>..] - ETA: 0s - loss: 0.3678 - acc: 0.8842 1793/1875 [===========================>..] - ETA: 0s - loss: 0.3658 - acc: 0.8848 1809/1875 [===========================>..] - ETA: 0s - loss: 0.3639 - acc: 0.8854 1825/1875 [============================>.] - ETA: 0s - loss: 0.3621 - acc: 0.8860 1841/1875 [============================>.] - ETA: 0s - loss: 0.3603 - acc: 0.8866 1858/1875 [============================>.] - ETA: 0s - loss: 0.3584 - acc: 0.8872 1875/1875 [==============================] - ETA: 0s - loss: 0.3565 - acc: 0.8878 1875/1875 [==============================] - 13s 5ms/step - loss: 0.3564 - acc: 0.8878 Epoch 2/5 1/1875 [..............................] - ETA: 12:55 - loss: 0.2051 - acc: 0.9375 11/1875 [..............................] - ETA: 9s - loss: 0.0763 - acc: 0.9769 22/1875 [..............................] - ETA: 9s - loss: 0.0737 - acc: 0.9791 33/1875 [..............................] - ETA: 8s - loss: 0.0669 - acc: 0.9817 44/1875 [..............................] - ETA: 8s - loss: 0.0643 - acc: 0.9826 54/1875 [..............................] - ETA: 8s - loss: 0.0624 - acc: 0.9831 65/1875 [>.............................] - ETA: 8s - loss: 0.0616 - acc: 0.9835 76/1875 [>.............................] - ETA: 8s - loss: 0.0613 - acc: 0.9836 86/1875 [>.............................] - ETA: 8s - loss: 0.0616 - acc: 0.9835 97/1875 [>.............................] - ETA: 8s - loss: 0.0618 - acc: 0.9835 108/1875 [>.............................] - ETA: 8s - loss: 0.0620 - acc: 0.9834 119/1875 [>.............................] - ETA: 8s - loss: 0.0622 - acc: 0.9833 129/1875 [=>............................] - ETA: 8s - loss: 0.0625 - acc: 0.9832 140/1875 [=>............................] - ETA: 8s - loss: 0.0626 - acc: 0.9831 151/1875 [=>............................] - ETA: 8s - loss: 0.0627 - acc: 0.9830 162/1875 [=>............................] - ETA: 8s - loss: 0.0628 - acc: 0.9829 172/1875 [=>............................] - ETA: 8s - loss: 0.0628 - acc: 0.9828 183/1875 [=>............................] - ETA: 8s - loss: 0.0627 - acc: 0.9828 194/1875 [==>...........................] - ETA: 8s - loss: 0.0626 - acc: 0.9828 205/1875 [==>...........................] - ETA: 8s - loss: 0.0625 - acc: 0.9827 215/1875 [==>...........................] - ETA: 8s - loss: 0.0624 - acc: 0.9827 225/1875 [==>...........................] - ETA: 8s - loss: 0.0623 - acc: 0.9827 235/1875 [==>...........................] - ETA: 8s - loss: 0.0622 - acc: 0.9827 245/1875 [==>...........................] - ETA: 7s - loss: 0.0621 - acc: 0.9827 255/1875 [===>..........................] - ETA: 7s - loss: 0.0620 - acc: 0.9827 266/1875 [===>..........................] - ETA: 7s - loss: 0.0619 - acc: 0.9827 277/1875 [===>..........................] - ETA: 7s - loss: 0.0618 - acc: 0.9827 288/1875 [===>..........................] - ETA: 7s - loss: 0.0617 - acc: 0.9827 299/1875 [===>..........................] - ETA: 7s - loss: 0.0616 - acc: 0.9827 310/1875 [===>..........................] - ETA: 7s - loss: 0.0615 - acc: 0.9827 321/1875 [====>.........................] - ETA: 7s - loss: 0.0614 - acc: 0.9827 332/1875 [====>.........................] - ETA: 7s - loss: 0.0613 - acc: 0.9827 342/1875 [====>.........................] - ETA: 7s - loss: 0.0612 - acc: 0.9827 353/1875 [====>.........................] - ETA: 7s - loss: 0.0611 - acc: 0.9827 364/1875 [====>.........................] - ETA: 7s - loss: 0.0610 - acc: 0.9827 375/1875 [=====>........................] - ETA: 7s - loss: 0.0610 - acc: 0.9827 386/1875 [=====>........................] - ETA: 7s - loss: 0.0609 - acc: 0.9827 398/1875 [=====>........................] - ETA: 7s - loss: 0.0608 - acc: 0.9826 410/1875 [=====>........................] - ETA: 7s - loss: 0.0607 - acc: 0.9826 421/1875 [=====>........................] - ETA: 7s - loss: 0.0606 - acc: 0.9826 432/1875 [=====>........................] - ETA: 6s - loss: 0.0605 - acc: 0.9827 444/1875 [======>.......................] - ETA: 6s - loss: 0.0603 - acc: 0.9827 456/1875 [======>.......................] - ETA: 6s - loss: 0.0602 - acc: 0.9827 467/1875 [======>.......................] - ETA: 6s - loss: 0.0601 - acc: 0.9827 478/1875 [======>.......................] - ETA: 6s - loss: 0.0600 - acc: 0.9827 489/1875 [======>.......................] - ETA: 6s - loss: 0.0599 - acc: 0.9827 500/1875 [=======>......................] - ETA: 6s - loss: 0.0598 - acc: 0.9827 511/1875 [=======>......................] - ETA: 6s - loss: 0.0597 - acc: 0.9828 521/1875 [=======>......................] - ETA: 6s - loss: 0.0596 - acc: 0.9828 532/1875 [=======>......................] - ETA: 6s - loss: 0.0595 - acc: 0.9828 543/1875 [=======>......................] - ETA: 6s - loss: 0.0594 - acc: 0.9828 554/1875 [=======>......................] - ETA: 6s - loss: 0.0593 - acc: 0.9828 565/1875 [========>.....................] - ETA: 6s - loss: 0.0592 - acc: 0.9828 577/1875 [========>.....................] - ETA: 6s - loss: 0.0592 - acc: 0.9828 589/1875 [========>.....................] - ETA: 6s - loss: 0.0591 - acc: 0.9828 601/1875 [========>.....................] - ETA: 6s - loss: 0.0590 - acc: 0.9828 612/1875 [========>.....................] - ETA: 6s - loss: 0.0589 - acc: 0.9829 624/1875 [========>.....................] - ETA: 5s - loss: 0.0588 - acc: 0.9829 636/1875 [=========>....................] - ETA: 5s - loss: 0.0587 - acc: 0.9829 648/1875 [=========>....................] - ETA: 5s - loss: 0.0586 - acc: 0.9829 659/1875 [=========>....................] - ETA: 5s - loss: 0.0585 - acc: 0.9829 670/1875 [=========>....................] - ETA: 5s - loss: 0.0584 - acc: 0.9830 681/1875 [=========>....................] - ETA: 5s - loss: 0.0583 - acc: 0.9830 693/1875 [==========>...................] - ETA: 5s - loss: 0.0582 - acc: 0.9830 704/1875 [==========>...................] - ETA: 5s - loss: 0.0581 - acc: 0.9830 715/1875 [==========>...................] - ETA: 5s - loss: 0.0581 - acc: 0.9830 727/1875 [==========>...................] - ETA: 5s - loss: 0.0580 - acc: 0.9831 739/1875 [==========>...................] - ETA: 5s - loss: 0.0579 - acc: 0.9831 750/1875 [===========>..................] - ETA: 5s - loss: 0.0578 - acc: 0.9831 761/1875 [===========>..................] - ETA: 5s - loss: 0.0578 - acc: 0.9831 772/1875 [===========>..................] - ETA: 5s - loss: 0.0577 - acc: 0.9831 783/1875 [===========>..................] - ETA: 5s - loss: 0.0576 - acc: 0.9832 794/1875 [===========>..................] - ETA: 5s - loss: 0.0576 - acc: 0.9832 806/1875 [===========>..................] - ETA: 5s - loss: 0.0575 - acc: 0.9832 817/1875 [============>.................] - ETA: 5s - loss: 0.0574 - acc: 0.9832 828/1875 [============>.................] - ETA: 4s - loss: 0.0574 - acc: 0.9832 839/1875 [============>.................] - ETA: 4s - loss: 0.0573 - acc: 0.9833 851/1875 [============>.................] - ETA: 4s - loss: 0.0572 - acc: 0.9833 863/1875 [============>.................] - ETA: 4s - loss: 0.0571 - acc: 0.9833 875/1875 [=============>................] - ETA: 4s - loss: 0.0571 - acc: 0.9833 886/1875 [=============>................] - ETA: 4s - loss: 0.0570 - acc: 0.9833 897/1875 [=============>................] - ETA: 4s - loss: 0.0570 - acc: 0.9833 909/1875 [=============>................] - ETA: 4s - loss: 0.0569 - acc: 0.9833 921/1875 [=============>................] - ETA: 4s - loss: 0.0568 - acc: 0.9834 932/1875 [=============>................] - ETA: 4s - loss: 0.0568 - acc: 0.9834 943/1875 [==============>...............] - ETA: 4s - loss: 0.0567 - acc: 0.9834 955/1875 [==============>...............] - ETA: 4s - loss: 0.0566 - acc: 0.9834 966/1875 [==============>...............] - ETA: 4s - loss: 0.0566 - acc: 0.9834 978/1875 [==============>...............] - ETA: 4s - loss: 0.0565 - acc: 0.9834 989/1875 [==============>...............] - ETA: 4s - loss: 0.0564 - acc: 0.9834 1001/1875 [===============>..............] - ETA: 4s - loss: 0.0564 - acc: 0.9835 1012/1875 [===============>..............] - ETA: 4s - loss: 0.0563 - acc: 0.9835 1023/1875 [===============>..............] - ETA: 3s - loss: 0.0562 - acc: 0.9835 1034/1875 [===============>..............] - ETA: 3s - loss: 0.0562 - acc: 0.9835 1046/1875 [===============>..............] - ETA: 3s - loss: 0.0561 - acc: 0.9835 1057/1875 [===============>..............] - ETA: 3s - loss: 0.0561 - acc: 0.9835 1068/1875 [================>.............] - ETA: 3s - loss: 0.0560 - acc: 0.9835 1079/1875 [================>.............] - ETA: 3s - loss: 0.0559 - acc: 0.9836 1090/1875 [================>.............] - ETA: 3s - loss: 0.0559 - acc: 0.9836 1101/1875 [================>.............] - ETA: 3s - loss: 0.0558 - acc: 0.9836 1113/1875 [================>.............] - ETA: 3s - loss: 0.0558 - acc: 0.9836 1125/1875 [=================>............] - ETA: 3s - loss: 0.0557 - acc: 0.9836 1137/1875 [=================>............] - ETA: 3s - loss: 0.0556 - acc: 0.9836 1148/1875 [=================>............] - ETA: 3s - loss: 0.0556 - acc: 0.9837 1159/1875 [=================>............] - ETA: 3s - loss: 0.0555 - acc: 0.9837 1171/1875 [=================>............] - ETA: 3s - loss: 0.0554 - acc: 0.9837 1183/1875 [=================>............] - ETA: 3s - loss: 0.0554 - acc: 0.9837 1195/1875 [==================>...........] - ETA: 3s - loss: 0.0553 - acc: 0.9837 1207/1875 [==================>...........] - ETA: 3s - loss: 0.0553 - acc: 0.9837 1219/1875 [==================>...........] - ETA: 3s - loss: 0.0552 - acc: 0.9837 1231/1875 [==================>...........] - ETA: 3s - loss: 0.0551 - acc: 0.9838 1242/1875 [==================>...........] - ETA: 2s - loss: 0.0551 - acc: 0.9838 1253/1875 [===================>..........] - ETA: 2s - loss: 0.0550 - acc: 0.9838 1264/1875 [===================>..........] - ETA: 2s - loss: 0.0550 - acc: 0.9838 1276/1875 [===================>..........] - ETA: 2s - loss: 0.0549 - acc: 0.9838 1287/1875 [===================>..........] - ETA: 2s - loss: 0.0549 - acc: 0.9838 1299/1875 [===================>..........] - ETA: 2s - loss: 0.0548 - acc: 0.9838 1311/1875 [===================>..........] - ETA: 2s - loss: 0.0548 - acc: 0.9839 1323/1875 [====================>.........] - ETA: 2s - loss: 0.0547 - acc: 0.9839 1334/1875 [====================>.........] - ETA: 2s - loss: 0.0547 - acc: 0.9839 1345/1875 [====================>.........] - ETA: 2s - loss: 0.0546 - acc: 0.9839 1357/1875 [====================>.........] - ETA: 2s - loss: 0.0546 - acc: 0.9839 1369/1875 [====================>.........] - ETA: 2s - loss: 0.0545 - acc: 0.9839 1380/1875 [=====================>........] - ETA: 2s - loss: 0.0545 - acc: 0.9839 1391/1875 [=====================>........] - ETA: 2s - loss: 0.0544 - acc: 0.9839 1402/1875 [=====================>........] - ETA: 2s - loss: 0.0544 - acc: 0.9839 1414/1875 [=====================>........] - ETA: 2s - loss: 0.0543 - acc: 0.9840 1426/1875 [=====================>........] - ETA: 2s - loss: 0.0543 - acc: 0.9840 1438/1875 [======================>.......] - ETA: 2s - loss: 0.0542 - acc: 0.9840 1450/1875 [======================>.......] - ETA: 1s - loss: 0.0542 - acc: 0.9840 1462/1875 [======================>.......] - ETA: 1s - loss: 0.0542 - acc: 0.9840 1474/1875 [======================>.......] - ETA: 1s - loss: 0.0541 - acc: 0.9840 1485/1875 [======================>.......] - ETA: 1s - loss: 0.0541 - acc: 0.9840 1496/1875 [======================>.......] - ETA: 1s - loss: 0.0540 - acc: 0.9840 1507/1875 [=======================>......] - ETA: 1s - loss: 0.0540 - acc: 0.9841 1518/1875 [=======================>......] - ETA: 1s - loss: 0.0539 - acc: 0.9841 1529/1875 [=======================>......] - ETA: 1s - loss: 0.0539 - acc: 0.9841 1539/1875 [=======================>......] - ETA: 1s - loss: 0.0539 - acc: 0.9841 1550/1875 [=======================>......] - ETA: 1s - loss: 0.0538 - acc: 0.9841 1562/1875 [=======================>......] - ETA: 1s - loss: 0.0538 - acc: 0.9841 1578/1875 [========================>.....] - ETA: 1s - loss: 0.0537 - acc: 0.9841 1595/1875 [========================>.....] - ETA: 1s - loss: 0.0536 - acc: 0.9841 1611/1875 [========================>.....] - ETA: 1s - loss: 0.0536 - acc: 0.9841 1627/1875 [=========================>....] - ETA: 1s - loss: 0.0535 - acc: 0.9842 1643/1875 [=========================>....] - ETA: 1s - loss: 0.0535 - acc: 0.9842 1660/1875 [=========================>....] - ETA: 0s - loss: 0.0534 - acc: 0.9842 1677/1875 [=========================>....] - ETA: 0s - loss: 0.0534 - acc: 0.9842 1693/1875 [==========================>...] - ETA: 0s - loss: 0.0533 - acc: 0.9842 1710/1875 [==========================>...] - ETA: 0s - loss: 0.0533 - acc: 0.9842 1727/1875 [==========================>...] - ETA: 0s - loss: 0.0532 - acc: 0.9842 1743/1875 [==========================>...] - ETA: 0s - loss: 0.0532 - acc: 0.9843 1759/1875 [===========================>..] - ETA: 0s - loss: 0.0531 - acc: 0.9843 1776/1875 [===========================>..] - ETA: 0s - loss: 0.0530 - acc: 0.9843 1793/1875 [===========================>..] - ETA: 0s - loss: 0.0530 - acc: 0.9843 1810/1875 [===========================>..] - ETA: 0s - loss: 0.0529 - acc: 0.9843 1827/1875 [============================>.] - ETA: 0s - loss: 0.0529 - acc: 0.9843 1844/1875 [============================>.] - ETA: 0s - loss: 0.0528 - acc: 0.9843 1861/1875 [============================>.] - ETA: 0s - loss: 0.0528 - acc: 0.9843 1875/1875 [==============================] - 9s 4ms/step - loss: 0.0528 - acc: 0.9843 Epoch 3/5 1/1875 [..............................] - ETA: 12:11 - loss: 0.0158 - acc: 1.0000 12/1875 [..............................] - ETA: 8s - loss: 0.0210 - acc: 0.9923 24/1875 [..............................] - ETA: 8s - loss: 0.0322 - acc: 0.9877 35/1875 [..............................] - ETA: 8s - loss: 0.0338 - acc: 0.9872 46/1875 [..............................] - ETA: 8s - loss: 0.0342 - acc: 0.9870 57/1875 [..............................] - ETA: 8s - loss: 0.0339 - acc: 0.9871 69/1875 [>.............................] - ETA: 8s - loss: 0.0330 - acc: 0.9875 81/1875 [>.............................] - ETA: 8s - loss: 0.0323 - acc: 0.9879 93/1875 [>.............................] - ETA: 8s - loss: 0.0318 - acc: 0.9882 105/1875 [>.............................] - ETA: 8s - loss: 0.0314 - acc: 0.9884 117/1875 [>.............................] - ETA: 7s - loss: 0.0313 - acc: 0.9886 129/1875 [=>............................] - ETA: 7s - loss: 0.0315 - acc: 0.9887 140/1875 [=>............................] - ETA: 7s - loss: 0.0315 - acc: 0.9888 152/1875 [=>............................] - ETA: 7s - loss: 0.0316 - acc: 0.9889 165/1875 [=>............................] - ETA: 7s - loss: 0.0316 - acc: 0.9890 177/1875 [=>............................] - ETA: 7s - loss: 0.0316 - acc: 0.9890 188/1875 [==>...........................] - ETA: 7s - loss: 0.0317 - acc: 0.9891 199/1875 [==>...........................] - ETA: 7s - loss: 0.0319 - acc: 0.9891 210/1875 [==>...........................] - ETA: 7s - loss: 0.0321 - acc: 0.9891 221/1875 [==>...........................] - ETA: 7s - loss: 0.0322 - acc: 0.9892 232/1875 [==>...........................] - ETA: 7s - loss: 0.0324 - acc: 0.9892 244/1875 [==>...........................] - ETA: 7s - loss: 0.0325 - acc: 0.9892 257/1875 [===>..........................] - ETA: 7s - loss: 0.0327 - acc: 0.9892 268/1875 [===>..........................] - ETA: 7s - loss: 0.0327 - acc: 0.9892 279/1875 [===>..........................] - ETA: 7s - loss: 0.0328 - acc: 0.9892 291/1875 [===>..........................] - ETA: 7s - loss: 0.0329 - acc: 0.9892 303/1875 [===>..........................] - ETA: 7s - loss: 0.0331 - acc: 0.9892 314/1875 [====>.........................] - ETA: 7s - loss: 0.0332 - acc: 0.9892 326/1875 [====>.........................] - ETA: 6s - loss: 0.0333 - acc: 0.9892 338/1875 [====>.........................] - ETA: 6s - loss: 0.0335 - acc: 0.9892 350/1875 [====>.........................] - ETA: 6s - loss: 0.0336 - acc: 0.9891 362/1875 [====>.........................] - ETA: 6s - loss: 0.0337 - acc: 0.9891 373/1875 [====>.........................] - ETA: 6s - loss: 0.0338 - acc: 0.9891 385/1875 [=====>........................] - ETA: 6s - loss: 0.0339 - acc: 0.9891 397/1875 [=====>........................] - ETA: 6s - loss: 0.0340 - acc: 0.9891 408/1875 [=====>........................] - ETA: 6s - loss: 0.0340 - acc: 0.9891 420/1875 [=====>........................] - ETA: 6s - loss: 0.0341 - acc: 0.9891 432/1875 [=====>........................] - ETA: 6s - loss: 0.0342 - acc: 0.9891 444/1875 [======>.......................] - ETA: 6s - loss: 0.0342 - acc: 0.9891 456/1875 [======>.......................] - ETA: 6s - loss: 0.0343 - acc: 0.9891 468/1875 [======>.......................] - ETA: 6s - loss: 0.0343 - acc: 0.9891 480/1875 [======>.......................] - ETA: 6s - loss: 0.0344 - acc: 0.9891 492/1875 [======>.......................] - ETA: 6s - loss: 0.0344 - acc: 0.9891 503/1875 [=======>......................] - ETA: 6s - loss: 0.0344 - acc: 0.9891 515/1875 [=======>......................] - ETA: 6s - loss: 0.0345 - acc: 0.9891 526/1875 [=======>......................] - ETA: 6s - loss: 0.0345 - acc: 0.9891 538/1875 [=======>......................] - ETA: 5s - loss: 0.0345 - acc: 0.9891 549/1875 [=======>......................] - ETA: 5s - loss: 0.0345 - acc: 0.9891 560/1875 [=======>......................] - ETA: 5s - loss: 0.0346 - acc: 0.9891 572/1875 [========>.....................] - ETA: 5s - loss: 0.0346 - acc: 0.9891 584/1875 [========>.....................] - ETA: 5s - loss: 0.0346 - acc: 0.9891 595/1875 [========>.....................] - ETA: 5s - loss: 0.0347 - acc: 0.9891 606/1875 [========>.....................] - ETA: 5s - loss: 0.0347 - acc: 0.9891 618/1875 [========>.....................] - ETA: 5s - loss: 0.0348 - acc: 0.9891 630/1875 [=========>....................] - ETA: 5s - loss: 0.0348 - acc: 0.9891 641/1875 [=========>....................] - ETA: 5s - loss: 0.0349 - acc: 0.9891 653/1875 [=========>....................] - ETA: 5s - loss: 0.0350 - acc: 0.9891 664/1875 [=========>....................] - ETA: 5s - loss: 0.0350 - acc: 0.9890 676/1875 [=========>....................] - ETA: 5s - loss: 0.0351 - acc: 0.9890 688/1875 [==========>...................] - ETA: 5s - loss: 0.0351 - acc: 0.9890 700/1875 [==========>...................] - ETA: 5s - loss: 0.0351 - acc: 0.9890 712/1875 [==========>...................] - ETA: 5s - loss: 0.0352 - acc: 0.9890 724/1875 [==========>...................] - ETA: 5s - loss: 0.0352 - acc: 0.9890 736/1875 [==========>...................] - ETA: 5s - loss: 0.0352 - acc: 0.9890 748/1875 [==========>...................] - ETA: 5s - loss: 0.0352 - acc: 0.9890 760/1875 [===========>..................] - ETA: 4s - loss: 0.0352 - acc: 0.9890 771/1875 [===========>..................] - ETA: 4s - loss: 0.0352 - acc: 0.9890 782/1875 [===========>..................] - ETA: 4s - loss: 0.0353 - acc: 0.9890 794/1875 [===========>..................] - ETA: 4s - loss: 0.0353 - acc: 0.9890 806/1875 [===========>..................] - ETA: 4s - loss: 0.0353 - acc: 0.9890 818/1875 [============>.................] - ETA: 4s - loss: 0.0353 - acc: 0.9890 830/1875 [============>.................] - ETA: 4s - loss: 0.0354 - acc: 0.9890 842/1875 [============>.................] - ETA: 4s - loss: 0.0354 - acc: 0.9890 854/1875 [============>.................] - ETA: 4s - loss: 0.0354 - acc: 0.9890 865/1875 [============>.................] - ETA: 4s - loss: 0.0354 - acc: 0.9890 876/1875 [=============>................] - ETA: 4s - loss: 0.0354 - acc: 0.9890 887/1875 [=============>................] - ETA: 4s - loss: 0.0354 - acc: 0.9890 899/1875 [=============>................] - ETA: 4s - loss: 0.0354 - acc: 0.9890 911/1875 [=============>................] - ETA: 4s - loss: 0.0355 - acc: 0.9890 923/1875 [=============>................] - ETA: 4s - loss: 0.0355 - acc: 0.9890 935/1875 [=============>................] - ETA: 4s - loss: 0.0355 - acc: 0.9890 947/1875 [==============>...............] - ETA: 4s - loss: 0.0355 - acc: 0.9890 958/1875 [==============>...............] - ETA: 4s - loss: 0.0355 - acc: 0.9890 970/1875 [==============>...............] - ETA: 4s - loss: 0.0355 - acc: 0.9890 982/1875 [==============>...............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 994/1875 [==============>...............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1005/1875 [===============>..............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1017/1875 [===============>..............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1029/1875 [===============>..............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1041/1875 [===============>..............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1053/1875 [===============>..............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1064/1875 [================>.............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1076/1875 [================>.............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1088/1875 [================>.............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1099/1875 [================>.............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1111/1875 [================>.............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1122/1875 [================>.............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1133/1875 [=================>............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1144/1875 [=================>............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1156/1875 [=================>............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1168/1875 [=================>............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1180/1875 [=================>............] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1192/1875 [==================>...........] - ETA: 3s - loss: 0.0355 - acc: 0.9890 1204/1875 [==================>...........] - ETA: 3s - loss: 0.0355 - acc: 0.9891 1216/1875 [==================>...........] - ETA: 2s - loss: 0.0355 - acc: 0.9891 1228/1875 [==================>...........] - ETA: 2s - loss: 0.0355 - acc: 0.9891 1240/1875 [==================>...........] - ETA: 2s - loss: 0.0355 - acc: 0.9891 1252/1875 [===================>..........] - ETA: 2s - loss: 0.0355 - acc: 0.9891 1263/1875 [===================>..........] - ETA: 2s - loss: 0.0355 - acc: 0.9891 1275/1875 [===================>..........] - ETA: 2s - loss: 0.0355 - acc: 0.9891 1287/1875 [===================>..........] - ETA: 2s - loss: 0.0355 - acc: 0.9891 1298/1875 [===================>..........] - ETA: 2s - loss: 0.0355 - acc: 0.9891 1310/1875 [===================>..........] - ETA: 2s - loss: 0.0354 - acc: 0.9891 1321/1875 [====================>.........] - ETA: 2s - loss: 0.0354 - acc: 0.9891 1332/1875 [====================>.........] - ETA: 2s - loss: 0.0354 - acc: 0.9891 1343/1875 [====================>.........] - ETA: 2s - loss: 0.0354 - acc: 0.9891 1354/1875 [====================>.........] - ETA: 2s - loss: 0.0354 - acc: 0.9891 1365/1875 [====================>.........] - ETA: 2s - loss: 0.0354 - acc: 0.9891 1377/1875 [=====================>........] - ETA: 2s - loss: 0.0354 - acc: 0.9891 1389/1875 [=====================>........] - ETA: 2s - loss: 0.0354 - acc: 0.9891 1401/1875 [=====================>........] - ETA: 2s - loss: 0.0354 - acc: 0.9891 1412/1875 [=====================>........] - ETA: 2s - loss: 0.0353 - acc: 0.9891 1423/1875 [=====================>........] - ETA: 2s - loss: 0.0353 - acc: 0.9891 1436/1875 [=====================>........] - ETA: 1s - loss: 0.0353 - acc: 0.9891 1448/1875 [======================>.......] - ETA: 1s - loss: 0.0353 - acc: 0.9891 1459/1875 [======================>.......] - ETA: 1s - loss: 0.0353 - acc: 0.9891 1470/1875 [======================>.......] - ETA: 1s - loss: 0.0353 - acc: 0.9891 1481/1875 [======================>.......] - ETA: 1s - loss: 0.0353 - acc: 0.9891 1493/1875 [======================>.......] - ETA: 1s - loss: 0.0353 - acc: 0.9891 1505/1875 [=======================>......] - ETA: 1s - loss: 0.0352 - acc: 0.9892 1516/1875 [=======================>......] - ETA: 1s - loss: 0.0352 - acc: 0.9892 1528/1875 [=======================>......] - ETA: 1s - loss: 0.0352 - acc: 0.9892 1540/1875 [=======================>......] - ETA: 1s - loss: 0.0352 - acc: 0.9892 1551/1875 [=======================>......] - ETA: 1s - loss: 0.0352 - acc: 0.9892 1563/1875 [========================>.....] - ETA: 1s - loss: 0.0352 - acc: 0.9892 1580/1875 [========================>.....] - ETA: 1s - loss: 0.0352 - acc: 0.9892 1596/1875 [========================>.....] - ETA: 1s - loss: 0.0351 - acc: 0.9892 1613/1875 [========================>.....] - ETA: 1s - loss: 0.0351 - acc: 0.9892 1630/1875 [=========================>....] - ETA: 1s - loss: 0.0351 - acc: 0.9892 1647/1875 [=========================>....] - ETA: 1s - loss: 0.0351 - acc: 0.9892 1664/1875 [=========================>....] - ETA: 0s - loss: 0.0350 - acc: 0.9892 1681/1875 [=========================>....] - ETA: 0s - loss: 0.0350 - acc: 0.9892 1698/1875 [==========================>...] - ETA: 0s - loss: 0.0350 - acc: 0.9892 1715/1875 [==========================>...] - ETA: 0s - loss: 0.0350 - acc: 0.9892 1731/1875 [==========================>...] - ETA: 0s - loss: 0.0350 - acc: 0.9892 1747/1875 [==========================>...] - ETA: 0s - loss: 0.0349 - acc: 0.9892 1763/1875 [===========================>..] - ETA: 0s - loss: 0.0349 - acc: 0.9892 1780/1875 [===========================>..] - ETA: 0s - loss: 0.0349 - acc: 0.9892 1796/1875 [===========================>..] - ETA: 0s - loss: 0.0349 - acc: 0.9893 1812/1875 [===========================>..] - ETA: 0s - loss: 0.0349 - acc: 0.9893 1829/1875 [============================>.] - ETA: 0s - loss: 0.0348 - acc: 0.9893 1846/1875 [============================>.] - ETA: 0s - loss: 0.0348 - acc: 0.9893 1862/1875 [============================>.] - ETA: 0s - loss: 0.0348 - acc: 0.9893 1875/1875 [==============================] - 8s 4ms/step - loss: 0.0348 - acc: 0.9893 Epoch 4/5 1/1875 [..............................] - ETA: 12:17 - loss: 0.0335 - acc: 1.0000 12/1875 [..............................] - ETA: 8s - loss: 0.0280 - acc: 0.9952 23/1875 [..............................] - ETA: 8s - loss: 0.0294 - acc: 0.9937 33/1875 [..............................] - ETA: 8s - loss: 0.0305 - acc: 0.9925 43/1875 [..............................] - ETA: 8s - loss: 0.0306 - acc: 0.9922 55/1875 [..............................] - ETA: 8s - loss: 0.0295 - acc: 0.9924 67/1875 [>.............................] - ETA: 8s - loss: 0.0292 - acc: 0.9923 79/1875 [>.............................] - ETA: 8s - loss: 0.0294 - acc: 0.9921 89/1875 [>.............................] - ETA: 8s - loss: 0.0293 - acc: 0.9921 100/1875 [>.............................] - ETA: 8s - loss: 0.0292 - acc: 0.9920 112/1875 [>.............................] - ETA: 8s - loss: 0.0291 - acc: 0.9919 124/1875 [>.............................] - ETA: 8s - loss: 0.0289 - acc: 0.9919 135/1875 [=>............................] - ETA: 8s - loss: 0.0288 - acc: 0.9918 146/1875 [=>............................] - ETA: 8s - loss: 0.0286 - acc: 0.9918 157/1875 [=>............................] - ETA: 8s - loss: 0.0284 - acc: 0.9917 168/1875 [=>............................] - ETA: 7s - loss: 0.0283 - acc: 0.9917 179/1875 [=>............................] - ETA: 7s - loss: 0.0283 - acc: 0.9916 190/1875 [==>...........................] - ETA: 7s - loss: 0.0282 - acc: 0.9916 202/1875 [==>...........................] - ETA: 7s - loss: 0.0281 - acc: 0.9916 214/1875 [==>...........................] - ETA: 7s - loss: 0.0280 - acc: 0.9916 226/1875 [==>...........................] - ETA: 7s - loss: 0.0280 - acc: 0.9915 238/1875 [==>...........................] - ETA: 7s - loss: 0.0279 - acc: 0.9915 249/1875 [==>...........................] - ETA: 7s - loss: 0.0279 - acc: 0.9915 260/1875 [===>..........................] - ETA: 7s - loss: 0.0279 - acc: 0.9915 271/1875 [===>..........................] - ETA: 7s - loss: 0.0278 - acc: 0.9915 283/1875 [===>..........................] - ETA: 7s - loss: 0.0278 - acc: 0.9914 295/1875 [===>..........................] - ETA: 7s - loss: 0.0277 - acc: 0.9914 306/1875 [===>..........................] - ETA: 7s - loss: 0.0277 - acc: 0.9914 317/1875 [====>.........................] - ETA: 7s - loss: 0.0277 - acc: 0.9914 329/1875 [====>.........................] - ETA: 7s - loss: 0.0277 - acc: 0.9914 341/1875 [====>.........................] - ETA: 7s - loss: 0.0277 - acc: 0.9914 352/1875 [====>.........................] - ETA: 7s - loss: 0.0277 - acc: 0.9914 363/1875 [====>.........................] - ETA: 6s - loss: 0.0276 - acc: 0.9914 375/1875 [=====>........................] - ETA: 6s - loss: 0.0276 - acc: 0.9914 387/1875 [=====>........................] - ETA: 6s - loss: 0.0276 - acc: 0.9914 398/1875 [=====>........................] - ETA: 6s - loss: 0.0277 - acc: 0.9913 409/1875 [=====>........................] - ETA: 6s - loss: 0.0277 - acc: 0.9913 421/1875 [=====>........................] - ETA: 6s - loss: 0.0277 - acc: 0.9913 433/1875 [=====>........................] - ETA: 6s - loss: 0.0278 - acc: 0.9913 444/1875 [======>.......................] - ETA: 6s - loss: 0.0278 - acc: 0.9913 457/1875 [======>.......................] - ETA: 6s - loss: 0.0278 - acc: 0.9913 469/1875 [======>.......................] - ETA: 6s - loss: 0.0278 - acc: 0.9913 481/1875 [======>.......................] - ETA: 6s - loss: 0.0279 - acc: 0.9912 493/1875 [======>.......................] - ETA: 6s - loss: 0.0279 - acc: 0.9912 505/1875 [=======>......................] - ETA: 6s - loss: 0.0279 - acc: 0.9912 518/1875 [=======>......................] - ETA: 6s - loss: 0.0280 - acc: 0.9912 530/1875 [=======>......................] - ETA: 6s - loss: 0.0280 - acc: 0.9912 542/1875 [=======>......................] - ETA: 6s - loss: 0.0280 - acc: 0.9912 553/1875 [=======>......................] - ETA: 5s - loss: 0.0280 - acc: 0.9912 565/1875 [========>.....................] - ETA: 5s - loss: 0.0280 - acc: 0.9912 576/1875 [========>.....................] - ETA: 5s - loss: 0.0280 - acc: 0.9912 588/1875 [========>.....................] - ETA: 5s - loss: 0.0280 - acc: 0.9912 600/1875 [========>.....................] - ETA: 5s - loss: 0.0280 - acc: 0.9912 612/1875 [========>.....................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 623/1875 [========>.....................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 634/1875 [=========>....................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 646/1875 [=========>....................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 658/1875 [=========>....................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 670/1875 [=========>....................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 682/1875 [=========>....................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 694/1875 [==========>...................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 706/1875 [==========>...................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 717/1875 [==========>...................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 729/1875 [==========>...................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 741/1875 [==========>...................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 753/1875 [===========>..................] - ETA: 5s - loss: 0.0280 - acc: 0.9911 764/1875 [===========>..................] - ETA: 5s - loss: 0.0279 - acc: 0.9911 776/1875 [===========>..................] - ETA: 4s - loss: 0.0279 - acc: 0.9911 788/1875 [===========>..................] - ETA: 4s - loss: 0.0279 - acc: 0.9911 800/1875 [===========>..................] - ETA: 4s - loss: 0.0279 - acc: 0.9911 811/1875 [===========>..................] - ETA: 4s - loss: 0.0279 - acc: 0.9911 823/1875 [============>.................] - ETA: 4s - loss: 0.0279 - acc: 0.9911 834/1875 [============>.................] - ETA: 4s - loss: 0.0278 - acc: 0.9911 846/1875 [============>.................] - ETA: 4s - loss: 0.0278 - acc: 0.9911 858/1875 [============>.................] - ETA: 4s - loss: 0.0278 - acc: 0.9911 870/1875 [============>.................] - ETA: 4s - loss: 0.0278 - acc: 0.9911 882/1875 [=============>................] - ETA: 4s - loss: 0.0278 - acc: 0.9911 894/1875 [=============>................] - ETA: 4s - loss: 0.0278 - acc: 0.9911 905/1875 [=============>................] - ETA: 4s - loss: 0.0278 - acc: 0.9911 917/1875 [=============>................] - ETA: 4s - loss: 0.0278 - acc: 0.9911 929/1875 [=============>................] - ETA: 4s - loss: 0.0278 - acc: 0.9911 941/1875 [==============>...............] - ETA: 4s - loss: 0.0278 - acc: 0.9911 952/1875 [==============>...............] - ETA: 4s - loss: 0.0278 - acc: 0.9911 964/1875 [==============>...............] - ETA: 4s - loss: 0.0278 - acc: 0.9911 976/1875 [==============>...............] - ETA: 4s - loss: 0.0278 - acc: 0.9911 987/1875 [==============>...............] - ETA: 3s - loss: 0.0278 - acc: 0.9911 999/1875 [==============>...............] - ETA: 3s - loss: 0.0278 - acc: 0.9911 1011/1875 [===============>..............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1023/1875 [===============>..............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1035/1875 [===============>..............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1047/1875 [===============>..............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1059/1875 [===============>..............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1072/1875 [================>.............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1084/1875 [================>.............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1095/1875 [================>.............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1107/1875 [================>.............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1119/1875 [================>.............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1131/1875 [=================>............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1142/1875 [=================>............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1153/1875 [=================>............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1165/1875 [=================>............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1176/1875 [=================>............] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1188/1875 [==================>...........] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1200/1875 [==================>...........] - ETA: 3s - loss: 0.0279 - acc: 0.9911 1211/1875 [==================>...........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1222/1875 [==================>...........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1233/1875 [==================>...........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1245/1875 [==================>...........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1257/1875 [===================>..........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1268/1875 [===================>..........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1280/1875 [===================>..........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1292/1875 [===================>..........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1304/1875 [===================>..........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1315/1875 [====================>.........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1327/1875 [====================>.........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1338/1875 [====================>.........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1350/1875 [====================>.........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1361/1875 [====================>.........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1373/1875 [====================>.........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1385/1875 [=====================>........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1397/1875 [=====================>........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1408/1875 [=====================>........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1419/1875 [=====================>........] - ETA: 2s - loss: 0.0279 - acc: 0.9911 1431/1875 [=====================>........] - ETA: 1s - loss: 0.0279 - acc: 0.9911 1443/1875 [======================>.......] - ETA: 1s - loss: 0.0279 - acc: 0.9911 1454/1875 [======================>.......] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1466/1875 [======================>.......] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1478/1875 [======================>.......] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1489/1875 [======================>.......] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1501/1875 [=======================>......] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1513/1875 [=======================>......] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1525/1875 [=======================>......] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1537/1875 [=======================>......] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1548/1875 [=======================>......] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1560/1875 [=======================>......] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1573/1875 [========================>.....] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1590/1875 [========================>.....] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1607/1875 [========================>.....] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1624/1875 [========================>.....] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1641/1875 [=========================>....] - ETA: 1s - loss: 0.0278 - acc: 0.9911 1658/1875 [=========================>....] - ETA: 0s - loss: 0.0277 - acc: 0.9911 1674/1875 [=========================>....] - ETA: 0s - loss: 0.0277 - acc: 0.9911 1691/1875 [==========================>...] - ETA: 0s - loss: 0.0277 - acc: 0.9911 1708/1875 [==========================>...] - ETA: 0s - loss: 0.0277 - acc: 0.9911 1724/1875 [==========================>...] - ETA: 0s - loss: 0.0277 - acc: 0.9911 1740/1875 [==========================>...] - ETA: 0s - loss: 0.0277 - acc: 0.9911 1757/1875 [===========================>..] - ETA: 0s - loss: 0.0277 - acc: 0.9911 1773/1875 [===========================>..] - ETA: 0s - loss: 0.0277 - acc: 0.9911 1788/1875 [===========================>..] - ETA: 0s - loss: 0.0277 - acc: 0.9911 1805/1875 [===========================>..] - ETA: 0s - loss: 0.0277 - acc: 0.9911 1822/1875 [============================>.] - ETA: 0s - loss: 0.0277 - acc: 0.9911 1839/1875 [============================>.] - ETA: 0s - loss: 0.0276 - acc: 0.9911 1855/1875 [============================>.] - ETA: 0s - loss: 0.0276 - acc: 0.9911 1871/1875 [============================>.] - ETA: 0s - loss: 0.0276 - acc: 0.9911 1875/1875 [==============================] - 8s 4ms/step - loss: 0.0276 - acc: 0.9911 Epoch 5/5 1/1875 [..............................] - ETA: 12:05 - loss: 0.0032 - acc: 1.0000 12/1875 [..............................] - ETA: 9s - loss: 0.0191 - acc: 0.9894 24/1875 [..............................] - ETA: 8s - loss: 0.0195 - acc: 0.9894 35/1875 [..............................] - ETA: 8s - loss: 0.0180 - acc: 0.9906 47/1875 [..............................] - ETA: 8s - loss: 0.0172 - acc: 0.9912 58/1875 [..............................] - ETA: 8s - loss: 0.0194 - acc: 0.9914 69/1875 [>.............................] - ETA: 8s - loss: 0.0210 - acc: 0.9914 81/1875 [>.............................] - ETA: 8s - loss: 0.0218 - acc: 0.9916 93/1875 [>.............................] - ETA: 8s - loss: 0.0221 - acc: 0.9918 105/1875 [>.............................] - ETA: 8s - loss: 0.0222 - acc: 0.9919 117/1875 [>.............................] - ETA: 7s - loss: 0.0222 - acc: 0.9921 129/1875 [=>............................] - ETA: 7s - loss: 0.0221 - acc: 0.9922 141/1875 [=>............................] - ETA: 7s - loss: 0.0221 - acc: 0.9924 153/1875 [=>............................] - ETA: 7s - loss: 0.0219 - acc: 0.9925 164/1875 [=>............................] - ETA: 7s - loss: 0.0218 - acc: 0.9926 175/1875 [=>............................] - ETA: 7s - loss: 0.0218 - acc: 0.9927 187/1875 [=>............................] - ETA: 7s - loss: 0.0217 - acc: 0.9928 199/1875 [==>...........................] - ETA: 7s - loss: 0.0217 - acc: 0.9929 210/1875 [==>...........................] - ETA: 7s - loss: 0.0216 - acc: 0.9930 221/1875 [==>...........................] - ETA: 7s - loss: 0.0215 - acc: 0.9931 232/1875 [==>...........................] - ETA: 7s - loss: 0.0214 - acc: 0.9932 244/1875 [==>...........................] - ETA: 7s - loss: 0.0213 - acc: 0.9932 255/1875 [===>..........................] - ETA: 7s - loss: 0.0212 - acc: 0.9933 266/1875 [===>..........................] - ETA: 7s - loss: 0.0212 - acc: 0.9933 278/1875 [===>..........................] - ETA: 7s - loss: 0.0212 - acc: 0.9934 290/1875 [===>..........................] - ETA: 7s - loss: 0.0212 - acc: 0.9934 302/1875 [===>..........................] - ETA: 7s - loss: 0.0212 - acc: 0.9934 314/1875 [====>.........................] - ETA: 7s - loss: 0.0212 - acc: 0.9935 326/1875 [====>.........................] - ETA: 6s - loss: 0.0212 - acc: 0.9935 338/1875 [====>.........................] - ETA: 6s - loss: 0.0212 - acc: 0.9935 349/1875 [====>.........................] - ETA: 6s - loss: 0.0212 - acc: 0.9936 360/1875 [====>.........................] - ETA: 6s - loss: 0.0212 - acc: 0.9936 372/1875 [====>.........................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 383/1875 [=====>........................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 394/1875 [=====>........................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 406/1875 [=====>........................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 418/1875 [=====>........................] - ETA: 6s - loss: 0.0210 - acc: 0.9936 430/1875 [=====>........................] - ETA: 6s - loss: 0.0210 - acc: 0.9936 441/1875 [======>.......................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 453/1875 [======>.......................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 465/1875 [======>.......................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 477/1875 [======>.......................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 488/1875 [======>.......................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 499/1875 [======>.......................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 510/1875 [=======>......................] - ETA: 6s - loss: 0.0210 - acc: 0.9936 522/1875 [=======>......................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 534/1875 [=======>......................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 545/1875 [=======>......................] - ETA: 6s - loss: 0.0211 - acc: 0.9936 556/1875 [=======>......................] - ETA: 5s - loss: 0.0211 - acc: 0.9936 567/1875 [========>.....................] - ETA: 5s - loss: 0.0211 - acc: 0.9936 578/1875 [========>.....................] - ETA: 5s - loss: 0.0212 - acc: 0.9936 589/1875 [========>.....................] - ETA: 5s - loss: 0.0212 - acc: 0.9935 601/1875 [========>.....................] - ETA: 5s - loss: 0.0212 - acc: 0.9935 611/1875 [========>.....................] - ETA: 5s - loss: 0.0212 - acc: 0.9935 622/1875 [========>.....................] - ETA: 5s - loss: 0.0213 - acc: 0.9935 634/1875 [=========>....................] - ETA: 5s - loss: 0.0213 - acc: 0.9935 646/1875 [=========>....................] - ETA: 5s - loss: 0.0213 - acc: 0.9935 658/1875 [=========>....................] - ETA: 5s - loss: 0.0213 - acc: 0.9935 669/1875 [=========>....................] - ETA: 5s - loss: 0.0214 - acc: 0.9935 680/1875 [=========>....................] - ETA: 5s - loss: 0.0214 - acc: 0.9935 691/1875 [==========>...................] - ETA: 5s - loss: 0.0214 - acc: 0.9935 702/1875 [==========>...................] - ETA: 5s - loss: 0.0214 - acc: 0.9935 713/1875 [==========>...................] - ETA: 5s - loss: 0.0214 - acc: 0.9935 725/1875 [==========>...................] - ETA: 5s - loss: 0.0214 - acc: 0.9935 736/1875 [==========>...................] - ETA: 5s - loss: 0.0214 - acc: 0.9935 747/1875 [==========>...................] - ETA: 5s - loss: 0.0214 - acc: 0.9935 759/1875 [===========>..................] - ETA: 5s - loss: 0.0214 - acc: 0.9935 772/1875 [===========>..................] - ETA: 5s - loss: 0.0214 - acc: 0.9935 784/1875 [===========>..................] - ETA: 4s - loss: 0.0214 - acc: 0.9935 796/1875 [===========>..................] - ETA: 4s - loss: 0.0214 - acc: 0.9935 808/1875 [===========>..................] - ETA: 4s - loss: 0.0214 - acc: 0.9935 820/1875 [============>.................] - ETA: 4s - loss: 0.0214 - acc: 0.9935 832/1875 [============>.................] - ETA: 4s - loss: 0.0214 - acc: 0.9935 844/1875 [============>.................] - ETA: 4s - loss: 0.0214 - acc: 0.9935 856/1875 [============>.................] - ETA: 4s - loss: 0.0214 - acc: 0.9935 868/1875 [============>.................] - ETA: 4s - loss: 0.0214 - acc: 0.9935 880/1875 [=============>................] - ETA: 4s - loss: 0.0214 - acc: 0.9935 890/1875 [=============>................] - ETA: 4s - loss: 0.0214 - acc: 0.9935 901/1875 [=============>................] - ETA: 4s - loss: 0.0213 - acc: 0.9935 912/1875 [=============>................] - ETA: 4s - loss: 0.0213 - acc: 0.9935 924/1875 [=============>................] - ETA: 4s - loss: 0.0213 - acc: 0.9935 935/1875 [=============>................] - ETA: 4s - loss: 0.0213 - acc: 0.9935 946/1875 [==============>...............] - ETA: 4s - loss: 0.0213 - acc: 0.9935 958/1875 [==============>...............] - ETA: 4s - loss: 0.0213 - acc: 0.9935 970/1875 [==============>...............] - ETA: 4s - loss: 0.0213 - acc: 0.9935 981/1875 [==============>...............] - ETA: 4s - loss: 0.0213 - acc: 0.9935 993/1875 [==============>...............] - ETA: 3s - loss: 0.0213 - acc: 0.9935 1004/1875 [===============>..............] - ETA: 3s - loss: 0.0213 - acc: 0.9935 1015/1875 [===============>..............] - ETA: 3s - loss: 0.0213 - acc: 0.9935 1026/1875 [===============>..............] - ETA: 3s - loss: 0.0214 - acc: 0.9935 1038/1875 [===============>..............] - ETA: 3s - loss: 0.0214 - acc: 0.9935 1050/1875 [===============>..............] - ETA: 3s - loss: 0.0214 - acc: 0.9935 1061/1875 [===============>..............] - ETA: 3s - loss: 0.0214 - acc: 0.9935 1072/1875 [================>.............] - ETA: 3s - loss: 0.0214 - acc: 0.9935 1083/1875 [================>.............] - ETA: 3s - loss: 0.0214 - acc: 0.9935 1095/1875 [================>.............] - ETA: 3s - loss: 0.0214 - acc: 0.9935 1106/1875 [================>.............] - ETA: 3s - loss: 0.0214 - acc: 0.9935 1117/1875 [================>.............] - ETA: 3s - loss: 0.0214 - acc: 0.9935 1128/1875 [=================>............] - ETA: 3s - loss: 0.0214 - acc: 0.9935 1140/1875 [=================>............] - ETA: 3s - loss: 0.0213 - acc: 0.9935 1152/1875 [=================>............] - ETA: 3s - loss: 0.0213 - acc: 0.9935 1162/1875 [=================>............] - ETA: 3s - loss: 0.0213 - acc: 0.9935 1174/1875 [=================>............] - ETA: 3s - loss: 0.0213 - acc: 0.9935 1186/1875 [=================>............] - ETA: 3s - loss: 0.0213 - acc: 0.9935 1197/1875 [==================>...........] - ETA: 3s - loss: 0.0213 - acc: 0.9935 1209/1875 [==================>...........] - ETA: 3s - loss: 0.0213 - acc: 0.9936 1220/1875 [==================>...........] - ETA: 2s - loss: 0.0213 - acc: 0.9936 1231/1875 [==================>...........] - ETA: 2s - loss: 0.0213 - acc: 0.9936 1243/1875 [==================>...........] - ETA: 2s - loss: 0.0213 - acc: 0.9936 1254/1875 [===================>..........] - ETA: 2s - loss: 0.0213 - acc: 0.9936 1266/1875 [===================>..........] - ETA: 2s - loss: 0.0213 - acc: 0.9936 1278/1875 [===================>..........] - ETA: 2s - loss: 0.0213 - acc: 0.9936 1290/1875 [===================>..........] - ETA: 2s - loss: 0.0213 - acc: 0.9936 1302/1875 [===================>..........] - ETA: 2s - loss: 0.0213 - acc: 0.9936 1313/1875 [====================>.........] - ETA: 2s - loss: 0.0213 - acc: 0.9936 1324/1875 [====================>.........] - ETA: 2s - loss: 0.0212 - acc: 0.9936 1335/1875 [====================>.........] - ETA: 2s - loss: 0.0212 - acc: 0.9936 1346/1875 [====================>.........] - ETA: 2s - loss: 0.0212 - acc: 0.9936 1358/1875 [====================>.........] - ETA: 2s - loss: 0.0212 - acc: 0.9936 1369/1875 [====================>.........] - ETA: 2s - loss: 0.0212 - acc: 0.9936 1381/1875 [=====================>........] - ETA: 2s - loss: 0.0212 - acc: 0.9936 1392/1875 [=====================>........] - ETA: 2s - loss: 0.0212 - acc: 0.9936 1403/1875 [=====================>........] - ETA: 2s - loss: 0.0212 - acc: 0.9936 1415/1875 [=====================>........] - ETA: 2s - loss: 0.0212 - acc: 0.9936 1426/1875 [=====================>........] - ETA: 2s - loss: 0.0212 - acc: 0.9936 1438/1875 [======================>.......] - ETA: 1s - loss: 0.0212 - acc: 0.9936 1450/1875 [======================>.......] - ETA: 1s - loss: 0.0212 - acc: 0.9936 1462/1875 [======================>.......] - ETA: 1s - loss: 0.0212 - acc: 0.9936 1474/1875 [======================>.......] - ETA: 1s - loss: 0.0212 - acc: 0.9936 1486/1875 [======================>.......] - ETA: 1s - loss: 0.0212 - acc: 0.9936 1498/1875 [======================>.......] - ETA: 1s - loss: 0.0212 - acc: 0.9936 1510/1875 [=======================>......] - ETA: 1s - loss: 0.0212 - acc: 0.9936 1521/1875 [=======================>......] - ETA: 1s - loss: 0.0212 - acc: 0.9936 1533/1875 [=======================>......] - ETA: 1s - loss: 0.0212 - acc: 0.9936 1545/1875 [=======================>......] - ETA: 1s - loss: 0.0211 - acc: 0.9937 1556/1875 [=======================>......] - ETA: 1s - loss: 0.0211 - acc: 0.9937 1566/1875 [========================>.....] - ETA: 1s - loss: 0.0211 - acc: 0.9937 1582/1875 [========================>.....] - ETA: 1s - loss: 0.0211 - acc: 0.9937 1598/1875 [========================>.....] - ETA: 1s - loss: 0.0211 - acc: 0.9937 1615/1875 [========================>.....] - ETA: 1s - loss: 0.0211 - acc: 0.9937 1631/1875 [=========================>....] - ETA: 1s - loss: 0.0211 - acc: 0.9937 1647/1875 [=========================>....] - ETA: 1s - loss: 0.0211 - acc: 0.9937 1664/1875 [=========================>....] - ETA: 0s - loss: 0.0211 - acc: 0.9937 1680/1875 [=========================>....] - ETA: 0s - loss: 0.0211 - acc: 0.9937 1696/1875 [==========================>...] - ETA: 0s - loss: 0.0211 - acc: 0.9937 1712/1875 [==========================>...] - ETA: 0s - loss: 0.0211 - acc: 0.9937 1729/1875 [==========================>...] - ETA: 0s - loss: 0.0211 - acc: 0.9937 1746/1875 [==========================>...] - ETA: 0s - loss: 0.0211 - acc: 0.9937 1762/1875 [===========================>..] - ETA: 0s - loss: 0.0211 - acc: 0.9937 1778/1875 [===========================>..] - ETA: 0s - loss: 0.0211 - acc: 0.9937 1795/1875 [===========================>..] - ETA: 0s - loss: 0.0211 - acc: 0.9937 1811/1875 [===========================>..] - ETA: 0s - loss: 0.0211 - acc: 0.9937 1828/1875 [============================>.] - ETA: 0s - loss: 0.0210 - acc: 0.9937 1845/1875 [============================>.] - ETA: 0s - loss: 0.0210 - acc: 0.9937 1862/1875 [============================>.] - ETA: 0s - loss: 0.0210 - acc: 0.9937 1875/1875 [==============================] - 8s 4ms/step - loss: 0.0210 - acc: 0.9937 <tensorflow.python.keras.callbacks.History at 0x7fa2901c7e48>
results = base_model.evaluate(test_dataset)
named_results = dict(zip(base_model.metrics_names, results))
print('\naccuracy:', named_results['acc'])
1/313 [..............................] - ETA: 2:34 - loss: 0.0051 - acc: 1.0000 19/313 [>.............................] - ETA: 0s - loss: 0.0402 - acc: 0.9934 38/313 [==>...........................] - ETA: 0s - loss: 0.0307 - acc: 0.9926 56/313 [====>.........................] - ETA: 0s - loss: 0.0306 - acc: 0.9927 75/313 [======>.......................] - ETA: 0s - loss: 0.0312 - acc: 0.9908 95/313 [========>.....................] - ETA: 0s - loss: 0.0339 - acc: 0.9898 114/313 [=========>....................] - ETA: 0s - loss: 0.0299 - acc: 0.9910 134/313 [===========>..................] - ETA: 0s - loss: 0.0293 - acc: 0.9909 154/313 [=============>................] - ETA: 0s - loss: 0.0347 - acc: 0.9903 173/313 [===============>..............] - ETA: 0s - loss: 0.0357 - acc: 0.9897 192/313 [=================>............] - ETA: 0s - loss: 0.0353 - acc: 0.9896 212/313 [===================>..........] - ETA: 0s - loss: 0.0354 - acc: 0.9898 232/313 [=====================>........] - ETA: 0s - loss: 0.0355 - acc: 0.9895 253/313 [=======================>......] - ETA: 0s - loss: 0.0363 - acc: 0.9895 274/313 [=========================>....] - ETA: 0s - loss: 0.0359 - acc: 0.9892 294/313 [===========================>..] - ETA: 0s - loss: 0.0353 - acc: 0.9892 313/313 [==============================] - ETA: 0s - loss: 0.0369 - acc: 0.9885 313/313 [==============================] - 1s 3ms/step - loss: 0.0369 - acc: 0.9885 accuracy: 0.9884999990463257
基本モデルがテストセットで 99% の精度を達成していることが分かります。下記の敵対的摂動下でのロバスト性では、モデルがどれだけロバストであるかを確認します。
敵対的正則化モデル
ここでは、NSL フレームワークを使用して、数行のコードで Keras モデルに敵対的トレーニングを組み込む方法を示します。基本モデルをラップして、トレーニング目的に敵対的正則化を含む新しい tf.Keras.Model を作成します。
まず最初に、ヘルパー関数 nsl.configs.make_adv_reg_config を使用して、関連するすべてのハイパーパラメータを含む構成オブジェクトを作成します。
adv_config = nsl.configs.make_adv_reg_config(
multiplier=HPARAMS.adv_multiplier,
adv_step_size=HPARAMS.adv_step_size,
adv_grad_norm=HPARAMS.adv_grad_norm
)
これで AdversarialRegularization を使って基本モデルをラップすることができます。ここでは、既存の基本モデル（base_model）を後で比較に使用できるように、新しい基本モデル（base_adv_model）を作成します。
返される adv_model は tf.keras.Model のオブジェクトであり、そのトレーニング目的には敵対的損失の正則化項を含みます。この損失を計算するためには、モデルは通常の正規入力（特徴 image）に加えてラベル情報（特徴 label）にアクセスする必要があります。この理由から、データセットの例は変換してタプルからディクショナリに戻します。そして label_keys パラメータを介してどの特徴がラベル情報を含んでいるかをモデルに伝えます。
base_adv_model = build_base_model(HPARAMS)
adv_model = nsl.keras.AdversarialRegularization(
base_adv_model,
label_keys=[LABEL_INPUT_NAME],
adv_config=adv_config
)
train_set_for_adv_model = train_dataset.map(convert_to_dictionaries)
test_set_for_adv_model = test_dataset.map(convert_to_dictionaries)
次に、敵対的正則化モデルをコンパイルしてトレーニングし、評価します。「損失ディクショナリに出力がありません」というような警告が出るかもしれませんが、これは adv_model が基本実装に依存せずトータルの損失を計算しているため、問題ありません。
adv_model.compile(optimizer='adam', loss='sparse_categorical_crossentropy',
metrics=['acc'])
adv_model.fit(train_set_for_adv_model, epochs=HPARAMS.epochs)
Epoch 1/5 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:AutoGraph could not transform <bound method Socket.send of <zmq.sugar.socket.Socket object at 0x7fa34d08c1e8>> and will run it as-is. Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: <cyfunction Socket.send at 0x7fa34afa72a0> is not a module, class, method, function, traceback, frame, or code object To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert WARNING:absl:Cannot perturb features dict_keys(['label'])WARNING:tensorflow:AutoGraph could not transform <bound method Socket.send of <zmq.sugar.socket.Socket object at 0x7fa34d08c1e8>> and will run it as-is. Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: <cyfunction Socket.send at 0x7fa34afa72a0> is not a module, class, method, function, traceback, frame, or code object To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert WARNING: AutoGraph could not transform <bound method Socket.send of <zmq.sugar.socket.Socket object at 0x7fa34d08c1e8>> and will run it as-is. Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output. Cause: <cyfunction Socket.send at 0x7fa34afa72a0> is not a module, class, method, function, traceback, frame, or code object To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 1/1875 [..............................] - ETA: 3:04:20 - loss: 2.7868 - sparse_categorical_crossentropy: 2.3079 - sparse_categorical_accuracy: 0.1250 - scaled_adversarial_loss: 0.4789 9/1875 [..............................] - ETA: 11s - loss: 2.7523 - sparse_categorical_crossentropy: 2.2779 - sparse_categorical_accuracy: 0.1698 - scaled_adversarial_loss: 0.4744 18/1875 [..............................] - ETA: 11s - loss: 2.6923 - sparse_categorical_crossentropy: 2.2222 - sparse_categorical_accuracy: 0.2411 - scaled_adversarial_loss: 0.4701 26/1875 [..............................] - ETA: 11s - loss: 2.6145 - sparse_categorical_crossentropy: 2.1470 - sparse_categorical_accuracy: 0.2858 - scaled_adversarial_loss: 0.4675 35/1875 [..............................] - ETA: 11s - loss: 2.5210 - sparse_categorical_crossentropy: 2.0538 - sparse_categorical_accuracy: 0.3299 - scaled_adversarial_loss: 0.4671 43/1875 [..............................] - ETA: 11s - loss: 2.4381 - sparse_categorical_crossentropy: 1.9704 - sparse_categorical_accuracy: 0.3638 - scaled_adversarial_loss: 0.4676 51/1875 [..............................] - ETA: 11s - loss: 2.3584 - sparse_categorical_crossentropy: 1.8909 - sparse_categorical_accuracy: 0.3936 - scaled_adversarial_loss: 0.4674 59/1875 [..............................] - ETA: 11s - loss: 2.2827 - sparse_categorical_crossentropy: 1.8165 - sparse_categorical_accuracy: 0.4203 - scaled_adversarial_loss: 0.4662 67/1875 [>.............................] - ETA: 11s - loss: 2.2116 - sparse_categorical_crossentropy: 1.7474 - sparse_categorical_accuracy: 0.4444 - scaled_adversarial_loss: 0.4641 75/1875 [>.............................] - ETA: 11s - loss: 2.1439 - sparse_categorical_crossentropy: 1.6825 - sparse_categorical_accuracy: 0.4669 - scaled_adversarial_loss: 0.4614 83/1875 [>.............................] - ETA: 11s - loss: 2.0809 - sparse_categorical_crossentropy: 1.6224 - sparse_categorical_accuracy: 0.4876 - scaled_adversarial_loss: 0.4585 91/1875 [>.............................] - ETA: 11s - loss: 2.0227 - sparse_categorical_crossentropy: 1.5671 - sparse_categorical_accuracy: 0.5064 - scaled_adversarial_loss: 0.4555 99/1875 [>.............................] - ETA: 11s - loss: 1.9691 - sparse_categorical_crossentropy: 1.5165 - sparse_categorical_accuracy: 0.5235 - scaled_adversarial_loss: 0.4527 107/1875 [>.............................] - ETA: 11s - loss: 1.9196 - sparse_categorical_crossentropy: 1.4698 - sparse_categorical_accuracy: 0.5392 - scaled_adversarial_loss: 0.4498 115/1875 [>.............................] - ETA: 11s - loss: 1.8737 - sparse_categorical_crossentropy: 1.4269 - sparse_categorical_accuracy: 0.5536 - scaled_adversarial_loss: 0.4468 123/1875 [>.............................] - ETA: 11s - loss: 1.8312 - sparse_categorical_crossentropy: 1.3872 - sparse_categorical_accuracy: 0.5668 - scaled_adversarial_loss: 0.4440 131/1875 [=>............................] - ETA: 11s - loss: 1.7912 - sparse_categorical_crossentropy: 1.3502 - sparse_categorical_accuracy: 0.5791 - scaled_adversarial_loss: 0.4411 140/1875 [=>............................] - ETA: 11s - loss: 1.7490 - sparse_categorical_crossentropy: 1.3113 - sparse_categorical_accuracy: 0.5919 - scaled_adversarial_loss: 0.4377 148/1875 [=>............................] - ETA: 11s - loss: 1.7135 - sparse_categorical_crossentropy: 1.2789 - sparse_categorical_accuracy: 0.6026 - scaled_adversarial_loss: 0.4347 156/1875 [=>............................] - ETA: 11s - loss: 1.6799 - sparse_categorical_crossentropy: 1.2483 - sparse_categorical_accuracy: 0.6126 - scaled_adversarial_loss: 0.4316 164/1875 [=>............................] - ETA: 11s - loss: 1.6482 - sparse_categorical_crossentropy: 1.2197 - sparse_categorical_accuracy: 0.6219 - scaled_adversarial_loss: 0.4285 172/1875 [=>............................] - ETA: 11s - loss: 1.6184 - sparse_categorical_crossentropy: 1.1928 - sparse_categorical_accuracy: 0.6306 - scaled_adversarial_loss: 0.4256 180/1875 [=>............................] - ETA: 11s - loss: 1.5900 - sparse_categorical_crossentropy: 1.1673 - sparse_categorical_accuracy: 0.6389 - scaled_adversarial_loss: 0.4226 188/1875 [==>...........................] - ETA: 11s - loss: 1.5630 - sparse_categorical_crossentropy: 1.1433 - sparse_categorical_accuracy: 0.6466 - scaled_adversarial_loss: 0.4198 196/1875 [==>...........................] - ETA: 10s - loss: 1.5375 - sparse_categorical_crossentropy: 1.1205 - sparse_categorical_accuracy: 0.6540 - scaled_adversarial_loss: 0.4170 205/1875 [==>...........................] - ETA: 10s - loss: 1.5101 - sparse_categorical_crossentropy: 1.0962 - sparse_categorical_accuracy: 0.6618 - scaled_adversarial_loss: 0.4139 213/1875 [==>...........................] - ETA: 10s - loss: 1.4870 - sparse_categorical_crossentropy: 1.0758 - sparse_categorical_accuracy: 0.6683 - scaled_adversarial_loss: 0.4112 221/1875 [==>...........................] - ETA: 10s - loss: 1.4648 - sparse_categorical_crossentropy: 1.0562 - sparse_categorical_accuracy: 0.6746 - scaled_adversarial_loss: 0.4086 229/1875 [==>...........................] - ETA: 10s - loss: 1.4436 - sparse_categorical_crossentropy: 1.0376 - sparse_categorical_accuracy: 0.6805 - scaled_adversarial_loss: 0.4060 237/1875 [==>...........................] - ETA: 10s - loss: 1.4234 - sparse_categorical_crossentropy: 1.0198 - sparse_categorical_accuracy: 0.6862 - scaled_adversarial_loss: 0.4035 245/1875 [==>...........................] - ETA: 10s - loss: 1.4038 - sparse_categorical_crossentropy: 1.0028 - sparse_categorical_accuracy: 0.6917 - scaled_adversarial_loss: 0.4010 253/1875 [===>..........................] - ETA: 10s - loss: 1.3851 - sparse_categorical_crossentropy: 0.9865 - sparse_categorical_accuracy: 0.6969 - scaled_adversarial_loss: 0.3986 261/1875 [===>..........................] - ETA: 10s - loss: 1.3671 - sparse_categorical_crossentropy: 0.9709 - sparse_categorical_accuracy: 0.7019 - scaled_adversarial_loss: 0.3962 269/1875 [===>..........................] - ETA: 10s - loss: 1.3498 - sparse_categorical_crossentropy: 0.9559 - sparse_categorical_accuracy: 0.7067 - scaled_adversarial_loss: 0.3939 277/1875 [===>..........................] - ETA: 10s - loss: 1.3332 - sparse_categorical_crossentropy: 0.9416 - sparse_categorical_accuracy: 0.7113 - scaled_adversarial_loss: 0.3916 285/1875 [===>..........................] - ETA: 10s - loss: 1.3171 - sparse_categorical_crossentropy: 0.9278 - sparse_categorical_accuracy: 0.7157 - scaled_adversarial_loss: 0.3894 294/1875 [===>..........................] - ETA: 10s - loss: 1.2997 - sparse_categorical_crossentropy: 0.9128 - sparse_categorical_accuracy: 0.7205 - scaled_adversarial_loss: 0.3869 303/1875 [===>..........................] - ETA: 10s - loss: 1.2829 - sparse_categorical_crossentropy: 0.8985 - sparse_categorical_accuracy: 0.7250 - scaled_adversarial_loss: 0.3844 310/1875 [===>..........................] - ETA: 10s - loss: 1.2703 - sparse_categorical_crossentropy: 0.8877 - sparse_categorical_accuracy: 0.7285 - scaled_adversarial_loss: 0.3826 318/1875 [====>.........................] - ETA: 10s - loss: 1.2563 - sparse_categorical_crossentropy: 0.8758 - sparse_categorical_accuracy: 0.7322 - scaled_adversarial_loss: 0.3805 326/1875 [====>.........................] - ETA: 10s - loss: 1.2428 - sparse_categorical_crossentropy: 0.8644 - sparse_categorical_accuracy: 0.7359 - scaled_adversarial_loss: 0.3785 334/1875 [====>.........................] - ETA: 9s - loss: 1.2297 - sparse_categorical_crossentropy: 0.8533 - sparse_categorical_accuracy: 0.7394 - scaled_adversarial_loss: 0.3764 342/1875 [====>.........................] - ETA: 9s - loss: 1.2170 - sparse_categorical_crossentropy: 0.8425 - sparse_categorical_accuracy: 0.7428 - scaled_adversarial_loss: 0.3745 350/1875 [====>.........................] - ETA: 9s - loss: 1.2047 - sparse_categorical_crossentropy: 0.8322 - sparse_categorical_accuracy: 0.7461 - scaled_adversarial_loss: 0.3726 358/1875 [====>.........................] - ETA: 9s - loss: 1.1928 - sparse_categorical_crossentropy: 0.8221 - sparse_categorical_accuracy: 0.7492 - scaled_adversarial_loss: 0.3707 366/1875 [====>.........................] - ETA: 9s - loss: 1.1812 - sparse_categorical_crossentropy: 0.8123 - sparse_categorical_accuracy: 0.7523 - scaled_adversarial_loss: 0.3688 374/1875 [====>.........................] - ETA: 9s - loss: 1.1699 - sparse_categorical_crossentropy: 0.8029 - sparse_categorical_accuracy: 0.7553 - scaled_adversarial_loss: 0.3670 382/1875 [=====>........................] - ETA: 9s - loss: 1.1589 - sparse_categorical_crossentropy: 0.7937 - sparse_categorical_accuracy: 0.7582 - scaled_adversarial_loss: 0.3652 390/1875 [=====>........................] - ETA: 9s - loss: 1.1483 - sparse_categorical_crossentropy: 0.7848 - sparse_categorical_accuracy: 0.7610 - scaled_adversarial_loss: 0.3635 398/1875 [=====>........................] - ETA: 9s - loss: 1.1379 - sparse_categorical_crossentropy: 0.7762 - sparse_categorical_accuracy: 0.7637 - scaled_adversarial_loss: 0.3618 406/1875 [=====>........................] - ETA: 9s - loss: 1.1279 - sparse_categorical_crossentropy: 0.7678 - sparse_categorical_accuracy: 0.7664 - scaled_adversarial_loss: 0.3601 414/1875 [=====>........................] - ETA: 9s - loss: 1.1180 - sparse_categorical_crossentropy: 0.7596 - sparse_categorical_accuracy: 0.7689 - scaled_adversarial_loss: 0.3584 423/1875 [=====>........................] - ETA: 9s - loss: 1.1074 - sparse_categorical_crossentropy: 0.7507 - sparse_categorical_accuracy: 0.7717 - scaled_adversarial_loss: 0.3566 431/1875 [=====>........................] - ETA: 9s - loss: 1.0981 - sparse_categorical_crossentropy: 0.7431 - sparse_categorical_accuracy: 0.7741 - scaled_adversarial_loss: 0.3550 439/1875 [======>.......................] - ETA: 9s - loss: 1.0891 - sparse_categorical_crossentropy: 0.7356 - sparse_categorical_accuracy: 0.7765 - scaled_adversarial_loss: 0.3535 448/1875 [======>.......................] - ETA: 9s - loss: 1.0792 - sparse_categorical_crossentropy: 0.7274 - sparse_categorical_accuracy: 0.7791 - scaled_adversarial_loss: 0.3518 457/1875 [======>.......................] - ETA: 9s - loss: 1.0695 - sparse_categorical_crossentropy: 0.7195 - sparse_categorical_accuracy: 0.7816 - scaled_adversarial_loss: 0.3501 465/1875 [======>.......................] - ETA: 9s - loss: 1.0612 - sparse_categorical_crossentropy: 0.7126 - sparse_categorical_accuracy: 0.7837 - scaled_adversarial_loss: 0.3486 473/1875 [======>.......................] - ETA: 9s - loss: 1.0530 - sparse_categorical_crossentropy: 0.7059 - sparse_categorical_accuracy: 0.7858 - scaled_adversarial_loss: 0.3471 482/1875 [======>.......................] - ETA: 9s - loss: 1.0441 - sparse_categorical_crossentropy: 0.6986 - sparse_categorical_accuracy: 0.7881 - scaled_adversarial_loss: 0.3455 491/1875 [======>.......................] - ETA: 8s - loss: 1.0354 - sparse_categorical_crossentropy: 0.6914 - sparse_categorical_accuracy: 0.7903 - scaled_adversarial_loss: 0.3439 499/1875 [======>.......................] - ETA: 8s - loss: 1.0278 - sparse_categorical_crossentropy: 0.6852 - sparse_categorical_accuracy: 0.7923 - scaled_adversarial_loss: 0.3426 507/1875 [=======>......................] - ETA: 8s - loss: 1.0204 - sparse_categorical_crossentropy: 0.6792 - sparse_categorical_accuracy: 0.7941 - scaled_adversarial_loss: 0.3412 516/1875 [=======>......................] - ETA: 8s - loss: 1.0123 - sparse_categorical_crossentropy: 0.6726 - sparse_categorical_accuracy: 0.7962 - scaled_adversarial_loss: 0.3397 524/1875 [=======>......................] - ETA: 8s - loss: 1.0053 - sparse_categorical_crossentropy: 0.6669 - sparse_categorical_accuracy: 0.7980 - scaled_adversarial_loss: 0.3384 532/1875 [=======>......................] - ETA: 8s - loss: 0.9984 - sparse_categorical_crossentropy: 0.6613 - sparse_categorical_accuracy: 0.7997 - scaled_adversarial_loss: 0.3371 541/1875 [=======>......................] - ETA: 8s - loss: 0.9908 - sparse_categorical_crossentropy: 0.6551 - sparse_categorical_accuracy: 0.8017 - scaled_adversarial_loss: 0.3357 549/1875 [=======>......................] - ETA: 8s - loss: 0.9842 - sparse_categorical_crossentropy: 0.6497 - sparse_categorical_accuracy: 0.8033 - scaled_adversarial_loss: 0.3345 557/1875 [=======>......................] - ETA: 8s - loss: 0.9777 - sparse_categorical_crossentropy: 0.6445 - sparse_categorical_accuracy: 0.8050 - scaled_adversarial_loss: 0.3332 565/1875 [========>.....................] - ETA: 8s - loss: 0.9714 - sparse_categorical_crossentropy: 0.6393 - sparse_categorical_accuracy: 0.8066 - scaled_adversarial_loss: 0.3320 573/1875 [========>.....................] - ETA: 8s - loss: 0.9652 - sparse_categorical_crossentropy: 0.6343 - sparse_categorical_accuracy: 0.8081 - scaled_adversarial_loss: 0.3309 582/1875 [========>.....................] - ETA: 8s - loss: 0.9583 - sparse_categorical_crossentropy: 0.6288 - sparse_categorical_accuracy: 0.8099 - scaled_adversarial_loss: 0.3295 590/1875 [========>.....................] - ETA: 8s - loss: 0.9523 - sparse_categorical_crossentropy: 0.6239 - sparse_categorical_accuracy: 0.8114 - scaled_adversarial_loss: 0.3284 598/1875 [========>.....................] - ETA: 8s - loss: 0.9465 - sparse_categorical_crossentropy: 0.6192 - sparse_categorical_accuracy: 0.8128 - scaled_adversarial_loss: 0.3272 606/1875 [========>.....................] - ETA: 8s - loss: 0.9407 - sparse_categorical_crossentropy: 0.6146 - sparse_categorical_accuracy: 0.8143 - scaled_adversarial_loss: 0.3261 614/1875 [========>.....................] - ETA: 8s - loss: 0.9351 - sparse_categorical_crossentropy: 0.6100 - sparse_categorical_accuracy: 0.8157 - scaled_adversarial_loss: 0.3250 622/1875 [========>.....................] - ETA: 8s - loss: 0.9295 - sparse_categorical_crossentropy: 0.6056 - sparse_categorical_accuracy: 0.8171 - scaled_adversarial_loss: 0.3239 630/1875 [=========>....................] - ETA: 8s - loss: 0.9241 - sparse_categorical_crossentropy: 0.6012 - sparse_categorical_accuracy: 0.8184 - scaled_adversarial_loss: 0.3229 639/1875 [=========>....................] - ETA: 7s - loss: 0.9181 - sparse_categorical_crossentropy: 0.5964 - sparse_categorical_accuracy: 0.8199 - scaled_adversarial_loss: 0.3217 648/1875 [=========>....................] - ETA: 7s - loss: 0.9122 - sparse_categorical_crossentropy: 0.5917 - sparse_categorical_accuracy: 0.8214 - scaled_adversarial_loss: 0.3205 656/1875 [=========>....................] - ETA: 7s - loss: 0.9070 - sparse_categorical_crossentropy: 0.5876 - sparse_categorical_accuracy: 0.8227 - scaled_adversarial_loss: 0.3194 664/1875 [=========>....................] - ETA: 7s - loss: 0.9019 - sparse_categorical_crossentropy: 0.5835 - sparse_categorical_accuracy: 0.8239 - scaled_adversarial_loss: 0.3184 672/1875 [=========>....................] - ETA: 7s - loss: 0.8969 - sparse_categorical_crossentropy: 0.5795 - sparse_categorical_accuracy: 0.8251 - scaled_adversarial_loss: 0.3174 680/1875 [=========>....................] - ETA: 7s - loss: 0.8920 - sparse_categorical_crossentropy: 0.5756 - sparse_categorical_accuracy: 0.8263 - scaled_adversarial_loss: 0.3164 688/1875 [==========>...................] - ETA: 7s - loss: 0.8872 - sparse_categorical_crossentropy: 0.5718 - sparse_categorical_accuracy: 0.8275 - scaled_adversarial_loss: 0.3154 696/1875 [==========>...................] - ETA: 7s - loss: 0.8825 - sparse_categorical_crossentropy: 0.5680 - sparse_categorical_accuracy: 0.8287 - scaled_adversarial_loss: 0.3144 704/1875 [==========>...................] - ETA: 7s - loss: 0.8778 - sparse_categorical_crossentropy: 0.5643 - sparse_categorical_accuracy: 0.8298 - scaled_adversarial_loss: 0.3135 712/1875 [==========>...................] - ETA: 7s - loss: 0.8732 - sparse_categorical_crossentropy: 0.5607 - sparse_categorical_accuracy: 0.8309 - scaled_adversarial_loss: 0.3125 720/1875 [==========>...................] - ETA: 7s - loss: 0.8687 - sparse_categorical_crossentropy: 0.5571 - sparse_categorical_accuracy: 0.8320 - scaled_adversarial_loss: 0.3115 728/1875 [==========>...................] - ETA: 7s - loss: 0.8642 - sparse_categorical_crossentropy: 0.5536 - sparse_categorical_accuracy: 0.8331 - scaled_adversarial_loss: 0.3106 736/1875 [==========>...................] - ETA: 7s - loss: 0.8598 - sparse_categorical_crossentropy: 0.5501 - sparse_categorical_accuracy: 0.8342 - scaled_adversarial_loss: 0.3097 745/1875 [==========>...................] - ETA: 7s - loss: 0.8550 - sparse_categorical_crossentropy: 0.5463 - sparse_categorical_accuracy: 0.8354 - scaled_adversarial_loss: 0.3087 753/1875 [===========>..................] - ETA: 7s - loss: 0.8507 - sparse_categorical_crossentropy: 0.5430 - sparse_categorical_accuracy: 0.8364 - scaled_adversarial_loss: 0.3078 761/1875 [===========>..................] - ETA: 7s - loss: 0.8465 - sparse_categorical_crossentropy: 0.5397 - sparse_categorical_accuracy: 0.8374 - scaled_adversarial_loss: 0.3069 770/1875 [===========>..................] - ETA: 7s - loss: 0.8419 - sparse_categorical_crossentropy: 0.5360 - sparse_categorical_accuracy: 0.8385 - scaled_adversarial_loss: 0.3059 778/1875 [===========>..................] - ETA: 7s - loss: 0.8378 - sparse_categorical_crossentropy: 0.5328 - sparse_categorical_accuracy: 0.8395 - scaled_adversarial_loss: 0.3050 786/1875 [===========>..................] - ETA: 7s - loss: 0.8339 - sparse_categorical_crossentropy: 0.5297 - sparse_categorical_accuracy: 0.8405 - scaled_adversarial_loss: 0.3041 794/1875 [===========>..................] - ETA: 6s - loss: 0.8299 - sparse_categorical_crossentropy: 0.5266 - sparse_categorical_accuracy: 0.8414 - scaled_adversarial_loss: 0.3033 802/1875 [===========>..................] - ETA: 6s - loss: 0.8261 - sparse_categorical_crossentropy: 0.5236 - sparse_categorical_accuracy: 0.8424 - scaled_adversarial_loss: 0.3024 810/1875 [===========>..................] - ETA: 6s - loss: 0.8222 - sparse_categorical_crossentropy: 0.5206 - sparse_categorical_accuracy: 0.8433 - scaled_adversarial_loss: 0.3016 818/1875 [============>.................] - ETA: 6s - loss: 0.8185 - sparse_categorical_crossentropy: 0.5177 - sparse_categorical_accuracy: 0.8442 - scaled_adversarial_loss: 0.3008 827/1875 [============>.................] - ETA: 6s - loss: 0.8143 - sparse_categorical_crossentropy: 0.5144 - sparse_categorical_accuracy: 0.8452 - scaled_adversarial_loss: 0.2999 836/1875 [============>.................] - ETA: 6s - loss: 0.8102 - sparse_categorical_crossentropy: 0.5112 - sparse_categorical_accuracy: 0.8462 - scaled_adversarial_loss: 0.2990 844/1875 [============>.................] - ETA: 6s - loss: 0.8066 - sparse_categorical_crossentropy: 0.5084 - sparse_categorical_accuracy: 0.8470 - scaled_adversarial_loss: 0.2982 852/1875 [============>.................] - ETA: 6s - loss: 0.8030 - sparse_categorical_crossentropy: 0.5056 - sparse_categorical_accuracy: 0.8479 - scaled_adversarial_loss: 0.2974 860/1875 [============>.................] - ETA: 6s - loss: 0.7995 - sparse_categorical_crossentropy: 0.5029 - sparse_categorical_accuracy: 0.8487 - scaled_adversarial_loss: 0.2966 868/1875 [============>.................] - ETA: 6s - loss: 0.7961 - sparse_categorical_crossentropy: 0.5002 - sparse_categorical_accuracy: 0.8495 - scaled_adversarial_loss: 0.2958 876/1875 [=============>................] - ETA: 6s - loss: 0.7927 - sparse_categorical_crossentropy: 0.4976 - sparse_categorical_accuracy: 0.8503 - scaled_adversarial_loss: 0.2951 884/1875 [=============>................] - ETA: 6s - loss: 0.7893 - sparse_categorical_crossentropy: 0.4950 - sparse_categorical_accuracy: 0.8511 - scaled_adversarial_loss: 0.2943 892/1875 [=============>................] - ETA: 6s - loss: 0.7860 - sparse_categorical_crossentropy: 0.4924 - sparse_categorical_accuracy: 0.8519 - scaled_adversarial_loss: 0.2936 900/1875 [=============>................] - ETA: 6s - loss: 0.7827 - sparse_categorical_crossentropy: 0.4899 - sparse_categorical_accuracy: 0.8527 - scaled_adversarial_loss: 0.2928 908/1875 [=============>................] - ETA: 6s - loss: 0.7795 - sparse_categorical_crossentropy: 0.4874 - sparse_categorical_accuracy: 0.8535 - scaled_adversarial_loss: 0.2921 916/1875 [=============>................] - ETA: 6s - loss: 0.7763 - sparse_categorical_crossentropy: 0.4849 - sparse_categorical_accuracy: 0.8542 - scaled_adversarial_loss: 0.2914 924/1875 [=============>................] - ETA: 6s - loss: 0.7731 - sparse_categorical_crossentropy: 0.4825 - sparse_categorical_accuracy: 0.8550 - scaled_adversarial_loss: 0.2907 932/1875 [=============>................] - ETA: 6s - loss: 0.7700 - sparse_categorical_crossentropy: 0.4801 - sparse_categorical_accuracy: 0.8557 - scaled_adversarial_loss: 0.2900 940/1875 [==============>...............] - ETA: 6s - loss: 0.7669 - sparse_categorical_crossentropy: 0.4777 - sparse_categorical_accuracy: 0.8564 - scaled_adversarial_loss: 0.2893 948/1875 [==============>...............] - ETA: 5s - loss: 0.7639 - sparse_categorical_crossentropy: 0.4753 - sparse_categorical_accuracy: 0.8572 - scaled_adversarial_loss: 0.2886 956/1875 [==============>...............] - ETA: 5s - loss: 0.7609 - sparse_categorical_crossentropy: 0.4730 - sparse_categorical_accuracy: 0.8579 - scaled_adversarial_loss: 0.2879 964/1875 [==============>...............] - ETA: 5s - loss: 0.7580 - sparse_categorical_crossentropy: 0.4708 - sparse_categorical_accuracy: 0.8586 - scaled_adversarial_loss: 0.2872 972/1875 [==============>...............] - ETA: 5s - loss: 0.7550 - sparse_categorical_crossentropy: 0.4685 - sparse_categorical_accuracy: 0.8592 - scaled_adversarial_loss: 0.2865 980/1875 [==============>...............] - ETA: 5s - loss: 0.7521 - sparse_categorical_crossentropy: 0.4663 - sparse_categorical_accuracy: 0.8599 - scaled_adversarial_loss: 0.2859 988/1875 [==============>...............] - ETA: 5s - loss: 0.7493 - sparse_categorical_crossentropy: 0.4641 - sparse_categorical_accuracy: 0.8606 - scaled_adversarial_loss: 0.2852 996/1875 [==============>...............] - ETA: 5s - loss: 0.7464 - sparse_categorical_crossentropy: 0.4619 - sparse_categorical_accuracy: 0.8613 - scaled_adversarial_loss: 0.2845 1004/1875 [===============>..............] - ETA: 5s - loss: 0.7436 - sparse_categorical_crossentropy: 0.4598 - sparse_categorical_accuracy: 0.8619 - scaled_adversarial_loss: 0.2839 1012/1875 [===============>..............] - ETA: 5s - loss: 0.7409 - sparse_categorical_crossentropy: 0.4577 - sparse_categorical_accuracy: 0.8626 - scaled_adversarial_loss: 0.2832 1020/1875 [===============>..............] - ETA: 5s - loss: 0.7381 - sparse_categorical_crossentropy: 0.4556 - sparse_categorical_accuracy: 0.8632 - scaled_adversarial_loss: 0.2826 1027/1875 [===============>..............] - ETA: 5s - loss: 0.7358 - sparse_categorical_crossentropy: 0.4538 - sparse_categorical_accuracy: 0.8637 - scaled_adversarial_loss: 0.2820 1035/1875 [===============>..............] - ETA: 5s - loss: 0.7331 - sparse_categorical_crossentropy: 0.4517 - sparse_categorical_accuracy: 0.8644 - scaled_adversarial_loss: 0.2814 1043/1875 [===============>..............] - ETA: 5s - loss: 0.7305 - sparse_categorical_crossentropy: 0.4497 - sparse_categorical_accuracy: 0.8650 - scaled_adversarial_loss: 0.2808 1051/1875 [===============>..............] - ETA: 5s - loss: 0.7278 - sparse_categorical_crossentropy: 0.4477 - sparse_categorical_accuracy: 0.8656 - scaled_adversarial_loss: 0.2801 1059/1875 [===============>..............] - ETA: 5s - loss: 0.7253 - sparse_categorical_crossentropy: 0.4457 - sparse_categorical_accuracy: 0.8662 - scaled_adversarial_loss: 0.2795 1067/1875 [================>.............] - ETA: 5s - loss: 0.7227 - sparse_categorical_crossentropy: 0.4438 - sparse_categorical_accuracy: 0.8668 - scaled_adversarial_loss: 0.2789 1075/1875 [================>.............] - ETA: 5s - loss: 0.7202 - sparse_categorical_crossentropy: 0.4419 - sparse_categorical_accuracy: 0.8674 - scaled_adversarial_loss: 0.2783 1083/1875 [================>.............] - ETA: 5s - loss: 0.7177 - sparse_categorical_crossentropy: 0.4400 - sparse_categorical_accuracy: 0.8679 - scaled_adversarial_loss: 0.2777 1091/1875 [================>.............] - ETA: 5s - loss: 0.7152 - sparse_categorical_crossentropy: 0.4381 - sparse_categorical_accuracy: 0.8685 - scaled_adversarial_loss: 0.2771 1099/1875 [================>.............] - ETA: 5s - loss: 0.7128 - sparse_categorical_crossentropy: 0.4362 - sparse_categorical_accuracy: 0.8691 - scaled_adversarial_loss: 0.2765 1107/1875 [================>.............] - ETA: 4s - loss: 0.7103 - sparse_categorical_crossentropy: 0.4344 - sparse_categorical_accuracy: 0.8696 - scaled_adversarial_loss: 0.2759 1115/1875 [================>.............] - ETA: 4s - loss: 0.7079 - sparse_categorical_crossentropy: 0.4326 - sparse_categorical_accuracy: 0.8702 - scaled_adversarial_loss: 0.2754 1123/1875 [================>.............] - ETA: 4s - loss: 0.7056 - sparse_categorical_crossentropy: 0.4308 - sparse_categorical_accuracy: 0.8707 - scaled_adversarial_loss: 0.2748 1131/1875 [=================>............] - ETA: 4s - loss: 0.7032 - sparse_categorical_crossentropy: 0.4290 - sparse_categorical_accuracy: 0.8713 - scaled_adversarial_loss: 0.2742 1139/1875 [=================>............] - ETA: 4s - loss: 0.7009 - sparse_categorical_crossentropy: 0.4273 - sparse_categorical_accuracy: 0.8718 - scaled_adversarial_loss: 0.2737 1147/1875 [=================>............] - ETA: 4s - loss: 0.6986 - sparse_categorical_crossentropy: 0.4255 - sparse_categorical_accuracy: 0.8723 - scaled_adversarial_loss: 0.2731 1155/1875 [=================>............] - ETA: 4s - loss: 0.6963 - sparse_categorical_crossentropy: 0.4238 - sparse_categorical_accuracy: 0.8729 - scaled_adversarial_loss: 0.2725 1163/1875 [=================>............] - ETA: 4s - loss: 0.6941 - sparse_categorical_crossentropy: 0.4221 - sparse_categorical_accuracy: 0.8734 - scaled_adversarial_loss: 0.2720 1171/1875 [=================>............] - ETA: 4s - loss: 0.6918 - sparse_categorical_crossentropy: 0.4204 - sparse_categorical_accuracy: 0.8739 - scaled_adversarial_loss: 0.2714 1179/1875 [=================>............] - ETA: 4s - loss: 0.6896 - sparse_categorical_crossentropy: 0.4187 - sparse_categorical_accuracy: 0.8744 - scaled_adversarial_loss: 0.2709 1187/1875 [=================>............] - ETA: 4s - loss: 0.6874 - sparse_categorical_crossentropy: 0.4171 - sparse_categorical_accuracy: 0.8749 - scaled_adversarial_loss: 0.2703 1195/1875 [==================>...........] - ETA: 4s - loss: 0.6853 - sparse_categorical_crossentropy: 0.4155 - sparse_categorical_accuracy: 0.8754 - scaled_adversarial_loss: 0.2698 1203/1875 [==================>...........] - ETA: 4s - loss: 0.6831 - sparse_categorical_crossentropy: 0.4138 - sparse_categorical_accuracy: 0.8759 - scaled_adversarial_loss: 0.2693 1211/1875 [==================>...........] - ETA: 4s - loss: 0.6810 - sparse_categorical_crossentropy: 0.4122 - sparse_categorical_accuracy: 0.8764 - scaled_adversarial_loss: 0.2688 1219/1875 [==================>...........] - ETA: 4s - loss: 0.6789 - sparse_categorical_crossentropy: 0.4107 - sparse_categorical_accuracy: 0.8769 - scaled_adversarial_loss: 0.2682 1227/1875 [==================>...........] - ETA: 4s - loss: 0.6768 - sparse_categorical_crossentropy: 0.4091 - sparse_categorical_accuracy: 0.8773 - scaled_adversarial_loss: 0.2677 1235/1875 [==================>...........] - ETA: 4s - loss: 0.6747 - sparse_categorical_crossentropy: 0.4076 - sparse_categorical_accuracy: 0.8778 - scaled_adversarial_loss: 0.2672 1243/1875 [==================>...........] - ETA: 4s - loss: 0.6727 - sparse_categorical_crossentropy: 0.4060 - sparse_categorical_accuracy: 0.8783 - scaled_adversarial_loss: 0.2667 1251/1875 [===================>..........] - ETA: 4s - loss: 0.6707 - sparse_categorical_crossentropy: 0.4045 - sparse_categorical_accuracy: 0.8787 - scaled_adversarial_loss: 0.2662 1260/1875 [===================>..........] - ETA: 3s - loss: 0.6684 - sparse_categorical_crossentropy: 0.4028 - sparse_categorical_accuracy: 0.8792 - scaled_adversarial_loss: 0.2656 1269/1875 [===================>..........] - ETA: 3s - loss: 0.6662 - sparse_categorical_crossentropy: 0.4012 - sparse_categorical_accuracy: 0.8797 - scaled_adversarial_loss: 0.2651 1277/1875 [===================>..........] - ETA: 3s - loss: 0.6642 - sparse_categorical_crossentropy: 0.3997 - sparse_categorical_accuracy: 0.8802 - scaled_adversarial_loss: 0.2646 1285/1875 [===================>..........] - ETA: 3s - loss: 0.6623 - sparse_categorical_crossentropy: 0.3982 - sparse_categorical_accuracy: 0.8806 - scaled_adversarial_loss: 0.2641 1293/1875 [===================>..........] - ETA: 3s - loss: 0.6604 - sparse_categorical_crossentropy: 0.3968 - sparse_categorical_accuracy: 0.8811 - scaled_adversarial_loss: 0.2636 1301/1875 [===================>..........] - ETA: 3s - loss: 0.6585 - sparse_categorical_crossentropy: 0.3954 - sparse_categorical_accuracy: 0.8815 - scaled_adversarial_loss: 0.2631 1309/1875 [===================>..........] - ETA: 3s - loss: 0.6566 - sparse_categorical_crossentropy: 0.3939 - sparse_categorical_accuracy: 0.8819 - scaled_adversarial_loss: 0.2626 1317/1875 [====================>.........] - ETA: 3s - loss: 0.6547 - sparse_categorical_crossentropy: 0.3925 - sparse_categorical_accuracy: 0.8823 - scaled_adversarial_loss: 0.2621 1325/1875 [====================>.........] - ETA: 3s - loss: 0.6528 - sparse_categorical_crossentropy: 0.3912 - sparse_categorical_accuracy: 0.8828 - scaled_adversarial_loss: 0.2617 1333/1875 [====================>.........] - ETA: 3s - loss: 0.6510 - sparse_categorical_crossentropy: 0.3898 - sparse_categorical_accuracy: 0.8832 - scaled_adversarial_loss: 0.2612 1341/1875 [====================>.........] - ETA: 3s - loss: 0.6491 - sparse_categorical_crossentropy: 0.3884 - sparse_categorical_accuracy: 0.8836 - scaled_adversarial_loss: 0.2607 1349/1875 [====================>.........] - ETA: 3s - loss: 0.6473 - sparse_categorical_crossentropy: 0.3871 - sparse_categorical_accuracy: 0.8840 - scaled_adversarial_loss: 0.2602 1357/1875 [====================>.........] - ETA: 3s - loss: 0.6455 - sparse_categorical_crossentropy: 0.3857 - sparse_categorical_accuracy: 0.8844 - scaled_adversarial_loss: 0.2598 1366/1875 [====================>.........] - ETA: 3s - loss: 0.6435 - sparse_categorical_crossentropy: 0.3842 - sparse_categorical_accuracy: 0.8849 - scaled_adversarial_loss: 0.2593 1374/1875 [====================>.........] - ETA: 3s - loss: 0.6417 - sparse_categorical_crossentropy: 0.3829 - sparse_categorical_accuracy: 0.8853 - scaled_adversarial_loss: 0.2588 1382/1875 [=====================>........] - ETA: 3s - loss: 0.6400 - sparse_categorical_crossentropy: 0.3816 - sparse_categorical_accuracy: 0.8856 - scaled_adversarial_loss: 0.2584 1390/1875 [=====================>........] - ETA: 3s - loss: 0.6383 - sparse_categorical_crossentropy: 0.3803 - sparse_categorical_accuracy: 0.8860 - scaled_adversarial_loss: 0.2579 1398/1875 [=====================>........] - ETA: 3s - loss: 0.6365 - sparse_categorical_crossentropy: 0.3791 - sparse_categorical_accuracy: 0.8864 - scaled_adversarial_loss: 0.2575 1406/1875 [=====================>........] - ETA: 3s - loss: 0.6348 - sparse_categorical_crossentropy: 0.3778 - sparse_categorical_accuracy: 0.8868 - scaled_adversarial_loss: 0.2570 1414/1875 [=====================>........] - ETA: 2s - loss: 0.6331 - sparse_categorical_crossentropy: 0.3765 - sparse_categorical_accuracy: 0.8872 - scaled_adversarial_loss: 0.2566 1422/1875 [=====================>........] - ETA: 2s - loss: 0.6315 - sparse_categorical_crossentropy: 0.3753 - sparse_categorical_accuracy: 0.8876 - scaled_adversarial_loss: 0.2561 1430/1875 [=====================>........] - ETA: 2s - loss: 0.6298 - sparse_categorical_crossentropy: 0.3741 - sparse_categorical_accuracy: 0.8879 - scaled_adversarial_loss: 0.2557 1437/1875 [=====================>........] - ETA: 2s - loss: 0.6283 - sparse_categorical_crossentropy: 0.3730 - sparse_categorical_accuracy: 0.8882 - scaled_adversarial_loss: 0.2553 1445/1875 [======================>.......] - ETA: 2s - loss: 0.6267 - sparse_categorical_crossentropy: 0.3718 - sparse_categorical_accuracy: 0.8886 - scaled_adversarial_loss: 0.2549 1453/1875 [======================>.......] - ETA: 2s - loss: 0.6251 - sparse_categorical_crossentropy: 0.3706 - sparse_categorical_accuracy: 0.8890 - scaled_adversarial_loss: 0.2545 1462/1875 [======================>.......] - ETA: 2s - loss: 0.6233 - sparse_categorical_crossentropy: 0.3693 - sparse_categorical_accuracy: 0.8894 - scaled_adversarial_loss: 0.2540 1470/1875 [======================>.......] - ETA: 2s - loss: 0.6217 - sparse_categorical_crossentropy: 0.3681 - sparse_categorical_accuracy: 0.8897 - scaled_adversarial_loss: 0.2536 1478/1875 [======================>.......] - ETA: 2s - loss: 0.6201 - sparse_categorical_crossentropy: 0.3669 - sparse_categorical_accuracy: 0.8901 - scaled_adversarial_loss: 0.2532 1486/1875 [======================>.......] - ETA: 2s - loss: 0.6185 - sparse_categorical_crossentropy: 0.3658 - sparse_categorical_accuracy: 0.8904 - scaled_adversarial_loss: 0.2528 1494/1875 [======================>.......] - ETA: 2s - loss: 0.6169 - sparse_categorical_crossentropy: 0.3646 - sparse_categorical_accuracy: 0.8908 - scaled_adversarial_loss: 0.2523 1502/1875 [=======================>......] - ETA: 2s - loss: 0.6154 - sparse_categorical_crossentropy: 0.3635 - sparse_categorical_accuracy: 0.8911 - scaled_adversarial_loss: 0.2519 1511/1875 [=======================>......] - ETA: 2s - loss: 0.6136 - sparse_categorical_crossentropy: 0.3622 - sparse_categorical_accuracy: 0.8915 - scaled_adversarial_loss: 0.2515 1519/1875 [=======================>......] - ETA: 2s - loss: 0.6121 - sparse_categorical_crossentropy: 0.3611 - sparse_categorical_accuracy: 0.8919 - scaled_adversarial_loss: 0.2511 1527/1875 [=======================>......] - ETA: 2s - loss: 0.6106 - sparse_categorical_crossentropy: 0.3599 - sparse_categorical_accuracy: 0.8922 - scaled_adversarial_loss: 0.2507 1536/1875 [=======================>......] - ETA: 2s - loss: 0.6089 - sparse_categorical_crossentropy: 0.3587 - sparse_categorical_accuracy: 0.8926 - scaled_adversarial_loss: 0.2502 1545/1875 [=======================>......] - ETA: 2s - loss: 0.6072 - sparse_categorical_crossentropy: 0.3575 - sparse_categorical_accuracy: 0.8929 - scaled_adversarial_loss: 0.2498 1553/1875 [=======================>......] - ETA: 2s - loss: 0.6057 - sparse_categorical_crossentropy: 0.3564 - sparse_categorical_accuracy: 0.8933 - scaled_adversarial_loss: 0.2494 1562/1875 [=======================>......] - ETA: 2s - loss: 0.6041 - sparse_categorical_crossentropy: 0.3552 - sparse_categorical_accuracy: 0.8936 - scaled_adversarial_loss: 0.2489 1571/1875 [========================>.....] - ETA: 1s - loss: 0.6025 - sparse_categorical_crossentropy: 0.3540 - sparse_categorical_accuracy: 0.8940 - scaled_adversarial_loss: 0.2485 1581/1875 [========================>.....] - ETA: 1s - loss: 0.6007 - sparse_categorical_crossentropy: 0.3527 - sparse_categorical_accuracy: 0.8944 - scaled_adversarial_loss: 0.2480 1591/1875 [========================>.....] - ETA: 1s - loss: 0.5989 - sparse_categorical_crossentropy: 0.3513 - sparse_categorical_accuracy: 0.8948 - scaled_adversarial_loss: 0.2475 1602/1875 [========================>.....] - ETA: 1s - loss: 0.5969 - sparse_categorical_crossentropy: 0.3499 - sparse_categorical_accuracy: 0.8952 - scaled_adversarial_loss: 0.2470 1612/1875 [========================>.....] - ETA: 1s - loss: 0.5952 - sparse_categorical_crossentropy: 0.3486 - sparse_categorical_accuracy: 0.8956 - scaled_adversarial_loss: 0.2465 1623/1875 [========================>.....] - ETA: 1s - loss: 0.5932 - sparse_categorical_crossentropy: 0.3473 - sparse_categorical_accuracy: 0.8960 - scaled_adversarial_loss: 0.2460 1634/1875 [=========================>....] - ETA: 1s - loss: 0.5914 - sparse_categorical_crossentropy: 0.3459 - sparse_categorical_accuracy: 0.8964 - scaled_adversarial_loss: 0.2455 1644/1875 [=========================>....] - ETA: 1s - loss: 0.5897 - sparse_categorical_crossentropy: 0.3446 - sparse_categorical_accuracy: 0.8968 - scaled_adversarial_loss: 0.2450 1654/1875 [=========================>....] - ETA: 1s - loss: 0.5880 - sparse_categorical_crossentropy: 0.3434 - sparse_categorical_accuracy: 0.8972 - scaled_adversarial_loss: 0.2446 1664/1875 [=========================>....] - ETA: 1s - loss: 0.5863 - sparse_categorical_crossentropy: 0.3422 - sparse_categorical_accuracy: 0.8976 - scaled_adversarial_loss: 0.2441 1674/1875 [=========================>....] - ETA: 1s - loss: 0.5846 - sparse_categorical_crossentropy: 0.3410 - sparse_categorical_accuracy: 0.8979 - scaled_adversarial_loss: 0.2436 1684/1875 [=========================>....] - ETA: 1s - loss: 0.5830 - sparse_categorical_crossentropy: 0.3398 - sparse_categorical_accuracy: 0.8983 - scaled_adversarial_loss: 0.2432 1695/1875 [==========================>...] - ETA: 1s - loss: 0.5812 - sparse_categorical_crossentropy: 0.3385 - sparse_categorical_accuracy: 0.8987 - scaled_adversarial_loss: 0.2427 1705/1875 [==========================>...] - ETA: 1s - loss: 0.5796 - sparse_categorical_crossentropy: 0.3373 - sparse_categorical_accuracy: 0.8990 - scaled_adversarial_loss: 0.2423 1715/1875 [==========================>...] - ETA: 1s - loss: 0.5780 - sparse_categorical_crossentropy: 0.3361 - sparse_categorical_accuracy: 0.8994 - scaled_adversarial_loss: 0.2418 1725/1875 [==========================>...] - ETA: 0s - loss: 0.5764 - sparse_categorical_crossentropy: 0.3350 - sparse_categorical_accuracy: 0.8997 - scaled_adversarial_loss: 0.2414 1735/1875 [==========================>...] - ETA: 0s - loss: 0.5748 - sparse_categorical_crossentropy: 0.3338 - sparse_categorical_accuracy: 0.9001 - scaled_adversarial_loss: 0.2409 1744/1875 [==========================>...] - ETA: 0s - loss: 0.5734 - sparse_categorical_crossentropy: 0.3328 - sparse_categorical_accuracy: 0.9004 - scaled_adversarial_loss: 0.2406 1753/1875 [===========================>..] - ETA: 0s - loss: 0.5720 - sparse_categorical_crossentropy: 0.3318 - sparse_categorical_accuracy: 0.9007 - scaled_adversarial_loss: 0.2402 1762/1875 [===========================>..] - ETA: 0s - loss: 0.5706 - sparse_categorical_crossentropy: 0.3308 - sparse_categorical_accuracy: 0.9010 - scaled_adversarial_loss: 0.2398 1770/1875 [===========================>..] - ETA: 0s - loss: 0.5694 - sparse_categorical_crossentropy: 0.3299 - sparse_categorical_accuracy: 0.9012 - scaled_adversarial_loss: 0.2394 1778/1875 [===========================>..] - ETA: 0s - loss: 0.5682 - sparse_categorical_crossentropy: 0.3291 - sparse_categorical_accuracy: 0.9015 - scaled_adversarial_loss: 0.2391 1787/1875 [===========================>..] - ETA: 0s - loss: 0.5668 - sparse_categorical_crossentropy: 0.3281 - sparse_categorical_accuracy: 0.9018 - scaled_adversarial_loss: 0.2387 1796/1875 [===========================>..] - ETA: 0s - loss: 0.5655 - sparse_categorical_crossentropy: 0.3271 - sparse_categorical_accuracy: 0.9021 - scaled_adversarial_loss: 0.2384 1806/1875 [===========================>..] - ETA: 0s - loss: 0.5640 - sparse_categorical_crossentropy: 0.3260 - sparse_categorical_accuracy: 0.9024 - scaled_adversarial_loss: 0.2379 1815/1875 [============================>.] - ETA: 0s - loss: 0.5627 - sparse_categorical_crossentropy: 0.3251 - sparse_categorical_accuracy: 0.9027 - scaled_adversarial_loss: 0.2376 1824/1875 [============================>.] - ETA: 0s - loss: 0.5613 - sparse_categorical_crossentropy: 0.3241 - sparse_categorical_accuracy: 0.9030 - scaled_adversarial_loss: 0.2372 1832/1875 [============================>.] - ETA: 0s - loss: 0.5602 - sparse_categorical_crossentropy: 0.3233 - sparse_categorical_accuracy: 0.9032 - scaled_adversarial_loss: 0.2369 1840/1875 [============================>.] - ETA: 0s - loss: 0.5590 - sparse_categorical_crossentropy: 0.3225 - sparse_categorical_accuracy: 0.9035 - scaled_adversarial_loss: 0.2366 1849/1875 [============================>.] - ETA: 0s - loss: 0.5577 - sparse_categorical_crossentropy: 0.3216 - sparse_categorical_accuracy: 0.9038 - scaled_adversarial_loss: 0.2362 1859/1875 [============================>.] - ETA: 0s - loss: 0.5563 - sparse_categorical_crossentropy: 0.3205 - sparse_categorical_accuracy: 0.9041 - scaled_adversarial_loss: 0.2358 1868/1875 [============================>.] - ETA: 0s - loss: 0.5551 - sparse_categorical_crossentropy: 0.3196 - sparse_categorical_accuracy: 0.9043 - scaled_adversarial_loss: 0.2354 1875/1875 [==============================] - 18s 6ms/step - loss: 0.5539 - sparse_categorical_crossentropy: 0.3188 - sparse_categorical_accuracy: 0.9046 - scaled_adversarial_loss: 0.2351 Epoch 2/5 1/1875 [..............................] - ETA: 13:28 - loss: 0.0638 - sparse_categorical_crossentropy: 0.0135 - sparse_categorical_accuracy: 1.0000 - scaled_adversarial_loss: 0.0503 8/1875 [..............................] - ETA: 14s - loss: 0.1413 - sparse_categorical_crossentropy: 0.0523 - sparse_categorical_accuracy: 0.9759 - scaled_adversarial_loss: 0.0890 15/1875 [..............................] - ETA: 14s - loss: 0.1574 - sparse_categorical_crossentropy: 0.0564 - sparse_categorical_accuracy: 0.9769 - scaled_adversarial_loss: 0.1011 22/1875 [..............................] - ETA: 13s - loss: 0.1618 - sparse_categorical_crossentropy: 0.0578 - sparse_categorical_accuracy: 0.9776 - scaled_adversarial_loss: 0.1040 30/1875 [..............................] - ETA: 13s - loss: 0.1597 - sparse_categorical_crossentropy: 0.0563 - sparse_categorical_accuracy: 0.9790 - scaled_adversarial_loss: 0.1034 37/1875 [..............................] - ETA: 13s - loss: 0.1580 - sparse_categorical_crossentropy: 0.0549 - sparse_categorical_accuracy: 0.9801 - scaled_adversarial_loss: 0.1031 44/1875 [..............................] - ETA: 13s - loss: 0.1561 - sparse_categorical_crossentropy: 0.0533 - sparse_categorical_accuracy: 0.9812 - scaled_adversarial_loss: 0.1028 51/1875 [..............................] - ETA: 13s - loss: 0.1560 - sparse_categorical_crossentropy: 0.0528 - sparse_categorical_accuracy: 0.9818 - scaled_adversarial_loss: 0.1032 58/1875 [..............................] - ETA: 13s - loss: 0.1563 - sparse_categorical_crossentropy: 0.0524 - sparse_categorical_accuracy: 0.9822 - scaled_adversarial_loss: 0.1038 66/1875 [>.............................] - ETA: 13s - loss: 0.1561 - sparse_categorical_crossentropy: 0.0520 - sparse_categorical_accuracy: 0.9826 - scaled_adversarial_loss: 0.1042 74/1875 [>.............................] - ETA: 13s - loss: 0.1556 - sparse_categorical_crossentropy: 0.0514 - sparse_categorical_accuracy: 0.9830 - scaled_adversarial_loss: 0.1042 82/1875 [>.............................] - ETA: 13s - loss: 0.1550 - sparse_categorical_crossentropy: 0.0509 - sparse_categorical_accuracy: 0.9833 - scaled_adversarial_loss: 0.1041 90/1875 [>.............................] - ETA: 12s - loss: 0.1546 - sparse_categorical_crossentropy: 0.0508 - sparse_categorical_accuracy: 0.9836 - scaled_adversarial_loss: 0.1038 97/1875 [>.............................] - ETA: 13s - loss: 0.1544 - sparse_categorical_crossentropy: 0.0508 - sparse_categorical_accuracy: 0.9838 - scaled_adversarial_loss: 0.1036 104/1875 [>.............................] - ETA: 12s - loss: 0.1543 - sparse_categorical_crossentropy: 0.0510 - sparse_categorical_accuracy: 0.9839 - scaled_adversarial_loss: 0.1034 112/1875 [>.............................] - ETA: 12s - loss: 0.1542 - sparse_categorical_crossentropy: 0.0510 - sparse_categorical_accuracy: 0.9840 - scaled_adversarial_loss: 0.1031 120/1875 [>.............................] - ETA: 12s - loss: 0.1538 - sparse_categorical_crossentropy: 0.0510 - sparse_categorical_accuracy: 0.9842 - scaled_adversarial_loss: 0.1028 128/1875 [=>............................] - ETA: 12s - loss: 0.1534 - sparse_categorical_crossentropy: 0.0509 - sparse_categorical_accuracy: 0.9843 - scaled_adversarial_loss: 0.1024 136/1875 [=>............................] - ETA: 12s - loss: 0.1530 - sparse_categorical_crossentropy: 0.0509 - sparse_categorical_accuracy: 0.9845 - scaled_adversarial_loss: 0.1021 143/1875 [=>............................] - ETA: 12s - loss: 0.1527 - sparse_categorical_crossentropy: 0.0509 - sparse_categorical_accuracy: 0.9845 - scaled_adversarial_loss: 0.1018 150/1875 [=>............................] - ETA: 12s - loss: 0.1524 - sparse_categorical_crossentropy: 0.0508 - sparse_categorical_accuracy: 0.9846 - scaled_adversarial_loss: 0.1016 157/1875 [=>............................] - ETA: 12s - loss: 0.1521 - sparse_categorical_crossentropy: 0.0508 - sparse_categorical_accuracy: 0.9847 - scaled_adversarial_loss: 0.1013 165/1875 [=>............................] - ETA: 12s - loss: 0.1517 - sparse_categorical_crossentropy: 0.0507 - sparse_categorical_accuracy: 0.9847 - scaled_adversarial_loss: 0.1009 173/1875 [=>............................] - ETA: 12s - loss: 0.1514 - sparse_categorical_crossentropy: 0.0507 - sparse_categorical_accuracy: 0.9848 - scaled_adversarial_loss: 0.1007 181/1875 [=>............................] - ETA: 12s - loss: 0.1511 - sparse_categorical_crossentropy: 0.0507 - sparse_categorical_accuracy: 0.9848 - scaled_adversarial_loss: 0.1004 189/1875 [==>...........................] - ETA: 12s - loss: 0.1508 - sparse_categorical_crossentropy: 0.0507 - sparse_categorical_accuracy: 0.9849 - scaled_adversarial_loss: 0.1002 196/1875 [==>...........................] - ETA: 12s - loss: 0.1505 - sparse_categorical_crossentropy: 0.0506 - sparse_categorical_accuracy: 0.9849 - scaled_adversarial_loss: 0.1000 203/1875 [==>...........................] - ETA: 12s - loss: 0.1503 - sparse_categorical_crossentropy: 0.0505 - sparse_categorical_accuracy: 0.9850 - scaled_adversarial_loss: 0.0998 211/1875 [==>...........................] - ETA: 12s - loss: 0.1499 - sparse_categorical_crossentropy: 0.0504 - sparse_categorical_accuracy: 0.9850 - scaled_adversarial_loss: 0.0995 219/1875 [==>...........................] - ETA: 11s - loss: 0.1496 - sparse_categorical_crossentropy: 0.0503 - sparse_categorical_accuracy: 0.9851 - scaled_adversarial_loss: 0.0993 226/1875 [==>...........................] - ETA: 11s - loss: 0.1493 - sparse_categorical_crossentropy: 0.0502 - sparse_categorical_accuracy: 0.9851 - scaled_adversarial_loss: 0.0991 233/1875 [==>...........................] - ETA: 11s - loss: 0.1490 - sparse_categorical_crossentropy: 0.0501 - sparse_categorical_accuracy: 0.9852 - scaled_adversarial_loss: 0.0989 240/1875 [==>...........................] - ETA: 11s - loss: 0.1486 - sparse_categorical_crossentropy: 0.0499 - sparse_categorical_accuracy: 0.9852 - scaled_adversarial_loss: 0.0987 247/1875 [==>...........................] - ETA: 11s - loss: 0.1484 - sparse_categorical_crossentropy: 0.0498 - sparse_categorical_accuracy: 0.9853 - scaled_adversarial_loss: 0.0985 254/1875 [===>..........................] - ETA: 11s - loss: 0.1481 - sparse_categorical_crossentropy: 0.0497 - sparse_categorical_accuracy: 0.9853 - scaled_adversarial_loss: 0.0984 262/1875 [===>..........................] - ETA: 11s - loss: 0.1478 - sparse_categorical_crossentropy: 0.0496 - sparse_categorical_accuracy: 0.9854 - scaled_adversarial_loss: 0.0982 269/1875 [===>..........................] - ETA: 11s - loss: 0.1475 - sparse_categorical_crossentropy: 0.0495 - sparse_categorical_accuracy: 0.9854 - scaled_adversarial_loss: 0.0980 277/1875 [===>..........................] - ETA: 11s - loss: 0.1473 - sparse_categorical_crossentropy: 0.0495 - sparse_categorical_accuracy: 0.9854 - scaled_adversarial_loss: 0.0979 285/1875 [===>..........................] - ETA: 11s - loss: 0.1471 - sparse_categorical_crossentropy: 0.0494 - sparse_categorical_accuracy: 0.9854 - scaled_adversarial_loss: 0.0977 292/1875 [===>..........................] - ETA: 11s - loss: 0.1469 - sparse_categorical_crossentropy: 0.0493 - sparse_categorical_accuracy: 0.9854 - scaled_adversarial_loss: 0.0976 299/1875 [===>..........................] - ETA: 11s - loss: 0.1467 - sparse_categorical_crossentropy: 0.0492 - sparse_categorical_accuracy: 0.9855 - scaled_adversarial_loss: 0.0974 307/1875 [===>..........................] - ETA: 11s - loss: 0.1465 - sparse_categorical_crossentropy: 0.0492 - sparse_categorical_accuracy: 0.9855 - scaled_adversarial_loss: 0.0973 315/1875 [====>.........................] - ETA: 11s - loss: 0.1463 - sparse_categorical_crossentropy: 0.0491 - sparse_categorical_accuracy: 0.9855 - scaled_adversarial_loss: 0.0972 322/1875 [====>.........................] - ETA: 11s - loss: 0.1461 - sparse_categorical_crossentropy: 0.0490 - sparse_categorical_accuracy: 0.9855 - scaled_adversarial_loss: 0.0971 330/1875 [====>.........................] - ETA: 11s - loss: 0.1459 - sparse_categorical_crossentropy: 0.0490 - sparse_categorical_accuracy: 0.9855 - scaled_adversarial_loss: 0.0969 337/1875 [====>.........................] - ETA: 11s - loss: 0.1457 - sparse_categorical_crossentropy: 0.0489 - sparse_categorical_accuracy: 0.9855 - scaled_adversarial_loss: 0.0968 344/1875 [====>.........................] - ETA: 11s - loss: 0.1456 - sparse_categorical_crossentropy: 0.0488 - sparse_categorical_accuracy: 0.9856 - scaled_adversarial_loss: 0.0967 351/1875 [====>.........................] - ETA: 11s - loss: 0.1454 - sparse_categorical_crossentropy: 0.0488 - sparse_categorical_accuracy: 0.9856 - scaled_adversarial_loss: 0.0966 359/1875 [====>.........................] - ETA: 10s - loss: 0.1453 - sparse_categorical_crossentropy: 0.0487 - sparse_categorical_accuracy: 0.9856 - scaled_adversarial_loss: 0.0965 366/1875 [====>.........................] - ETA: 10s - loss: 0.1451 - sparse_categorical_crossentropy: 0.0487 - sparse_categorical_accuracy: 0.9856 - scaled_adversarial_loss: 0.0965 374/1875 [====>.........................] - ETA: 10s - loss: 0.1450 - sparse_categorical_crossentropy: 0.0486 - sparse_categorical_accuracy: 0.9856 - scaled_adversarial_loss: 0.0963 381/1875 [=====>........................] - ETA: 10s - loss: 0.1448 - sparse_categorical_crossentropy: 0.0486 - sparse_categorical_accuracy: 0.9856 - scaled_adversarial_loss: 0.0963 388/1875 [=====>........................] - ETA: 10s - loss: 0.1447 - sparse_categorical_crossentropy: 0.0485 - sparse_categorical_accuracy: 0.9856 - scaled_adversarial_loss: 0.0962 395/1875 [=====>........................] - ETA: 10s - loss: 0.1446 - sparse_categorical_crossentropy: 0.0485 - sparse_categorical_accuracy: 0.9856 - scaled_adversarial_loss: 0.0961 403/1875 [=====>........................] - ETA: 10s - loss: 0.1444 - sparse_categorical_crossentropy: 0.0485 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0960 411/1875 [=====>........................] - ETA: 10s - loss: 0.1443 - sparse_categorical_crossentropy: 0.0484 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0959 419/1875 [=====>........................] - ETA: 10s - loss: 0.1441 - sparse_categorical_crossentropy: 0.0484 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0958 427/1875 [=====>........................] - ETA: 10s - loss: 0.1440 - sparse_categorical_crossentropy: 0.0483 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0957 434/1875 [=====>........................] - ETA: 10s - loss: 0.1439 - sparse_categorical_crossentropy: 0.0483 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0956 441/1875 [======>.......................] - ETA: 10s - loss: 0.1438 - sparse_categorical_crossentropy: 0.0483 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0955 448/1875 [======>.......................] - ETA: 10s - loss: 0.1437 - sparse_categorical_crossentropy: 0.0482 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0955 455/1875 [======>.......................] - ETA: 10s - loss: 0.1436 - sparse_categorical_crossentropy: 0.0482 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0954 462/1875 [======>.......................] - ETA: 10s - loss: 0.1435 - sparse_categorical_crossentropy: 0.0482 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0953 469/1875 [======>.......................] - ETA: 10s - loss: 0.1434 - sparse_categorical_crossentropy: 0.0481 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0953 477/1875 [======>.......................] - ETA: 10s - loss: 0.1433 - sparse_categorical_crossentropy: 0.0481 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0952 484/1875 [======>.......................] - ETA: 10s - loss: 0.1432 - sparse_categorical_crossentropy: 0.0481 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0951 491/1875 [======>.......................] - ETA: 10s - loss: 0.1431 - sparse_categorical_crossentropy: 0.0480 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0950 498/1875 [======>.......................] - ETA: 10s - loss: 0.1429 - sparse_categorical_crossentropy: 0.0480 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0950 505/1875 [=======>......................] - ETA: 9s - loss: 0.1428 - sparse_categorical_crossentropy: 0.0480 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0949 512/1875 [=======>......................] - ETA: 9s - loss: 0.1427 - sparse_categorical_crossentropy: 0.0479 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0948 519/1875 [=======>......................] - ETA: 9s - loss: 0.1426 - sparse_categorical_crossentropy: 0.0479 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0947 526/1875 [=======>......................] - ETA: 9s - loss: 0.1425 - sparse_categorical_crossentropy: 0.0479 - sparse_categorical_accuracy: 0.9857 - scaled_adversarial_loss: 0.0947 533/1875 [=======>......................] - ETA: 9s - loss: 0.1424 - sparse_categorical_crossentropy: 0.0478 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0946 540/1875 [=======>......................] - ETA: 9s - loss: 0.1424 - sparse_categorical_crossentropy: 0.0478 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0946 547/1875 [=======>......................] - ETA: 9s - loss: 0.1423 - sparse_categorical_crossentropy: 0.0478 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0945 554/1875 [=======>......................] - ETA: 9s - loss: 0.1422 - sparse_categorical_crossentropy: 0.0477 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0944 561/1875 [=======>......................] - ETA: 9s - loss: 0.1421 - sparse_categorical_crossentropy: 0.0477 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0944 569/1875 [========>.....................] - ETA: 9s - loss: 0.1419 - sparse_categorical_crossentropy: 0.0476 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0943 576/1875 [========>.....................] - ETA: 9s - loss: 0.1418 - sparse_categorical_crossentropy: 0.0476 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0942 583/1875 [========>.....................] - ETA: 9s - loss: 0.1417 - sparse_categorical_crossentropy: 0.0476 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0942 590/1875 [========>.....................] - ETA: 9s - loss: 0.1416 - sparse_categorical_crossentropy: 0.0475 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0941 598/1875 [========>.....................] - ETA: 9s - loss: 0.1415 - sparse_categorical_crossentropy: 0.0475 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0940 606/1875 [========>.....................] - ETA: 9s - loss: 0.1414 - sparse_categorical_crossentropy: 0.0474 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0939 614/1875 [========>.....................] - ETA: 9s - loss: 0.1412 - sparse_categorical_crossentropy: 0.0474 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0939 621/1875 [========>.....................] - ETA: 9s - loss: 0.1411 - sparse_categorical_crossentropy: 0.0473 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0938 628/1875 [=========>....................] - ETA: 9s - loss: 0.1410 - sparse_categorical_crossentropy: 0.0473 - sparse_categorical_accuracy: 0.9858 - scaled_adversarial_loss: 0.0937 635/1875 [=========>....................] - ETA: 9s - loss: 0.1409 - sparse_categorical_crossentropy: 0.0473 - sparse_categorical_accuracy: 0.9859 - scaled_adversarial_loss: 0.0937 643/1875 [=========>....................] - ETA: 8s - loss: 0.1408 - sparse_categorical_crossentropy: 0.0472 - sparse_categorical_accuracy: 0.9859 - scaled_adversarial_loss: 0.0936 650/1875 [=========>....................] - ETA: 8s - loss: 0.1407 - sparse_categorical_crossentropy: 0.0472 - sparse_categorical_accuracy: 0.9859 - scaled_adversarial_loss: 0.0935 658/1875 [=========>....................] - ETA: 8s - loss: 0.1406 - sparse_categorical_crossentropy: 0.0471 - sparse_categorical_accuracy: 0.9859 - scaled_adversarial_loss: 0.0934 666/1875 [=========>....................] - ETA: 8s - loss: 0.1405 - sparse_categorical_crossentropy: 0.0471 - sparse_categorical_accuracy: 0.9859 - scaled_adversarial_loss: 0.0934 673/1875 [=========>....................] - ETA: 8s - loss: 0.1404 - sparse_categorical_crossentropy: 0.0471 - sparse_categorical_accuracy: 0.9859 - scaled_adversarial_loss: 0.0933 680/1875 [=========>....................] - ETA: 8s - loss: 0.1403 - sparse_categorical_crossentropy: 0.0470 - sparse_categorical_accuracy: 0.9859 - scaled_adversarial_loss: 0.0932 687/1875 [=========>....................] - ETA: 8s - loss: 0.1402 - sparse_categorical_crossentropy: 0.0470 - sparse_categorical_accuracy: 0.9859 - scaled_adversarial_loss: 0.0932 695/1875 [==========>...................] - ETA: 8s - loss: 0.1400 - sparse_categorical_crossentropy: 0.0469 - sparse_categorical_accuracy: 0.9859 - scaled_adversarial_loss: 0.0931 702/1875 [==========>...................] - ETA: 8s - loss: 0.1400 - sparse_categorical_crossentropy: 0.0469 - sparse_categorical_accuracy: 0.9860 - scaled_adversarial_loss: 0.0930 710/1875 [==========>...................] - ETA: 8s - loss: 0.1398 - sparse_categorical_crossentropy: 0.0469 - sparse_categorical_accuracy: 0.9860 - scaled_adversarial_loss: 0.0930 718/1875 [==========>...................] - ETA: 8s - loss: 0.1397 - sparse_categorical_crossentropy: 0.0468 - sparse_categorical_accuracy: 0.9860 - scaled_adversarial_loss: 0.0929 725/1875 [==========>...................] - ETA: 8s - loss: 0.1396 - sparse_categorical_crossentropy: 0.0468 - sparse_categorical_accuracy: 0.9860 - scaled_adversarial_loss: 0.0928 732/1875 [==========>...................] - ETA: 8s - loss: 0.1395 - sparse_categorical_crossentropy: 0.0468 - sparse_categorical_accuracy: 0.9860 - scaled_adversarial_loss: 0.0928 739/1875 [==========>...................] - ETA: 8s - loss: 0.1394 - sparse_categorical_crossentropy: 0.0467 - sparse_categorical_accuracy: 0.9860 - scaled_adversarial_loss: 0.0927 746/1875 [==========>...................] - ETA: 8s - loss: 0.1394 - sparse_categorical_crossentropy: 0.0467 - sparse_categorical_accuracy: 0.9860 - scaled_adversarial_loss: 0.0927 753/1875 [===========>..................] - ETA: 8s - loss: 0.1393 - sparse_categorical_crossentropy: 0.0466 - sparse_categorical_accuracy: 0.9860 - scaled_adversarial_loss: 0.0926 760/1875 [===========>..................] - ETA: 8s - loss: 0.1392 - sparse_categorical_crossentropy: 0.0466 - sparse_categorical_accuracy: 0.9861 - scaled_adversarial_loss: 0.0926 767/1875 [===========>..................] - ETA: 8s - loss: 0.1391 - sparse_categorical_crossentropy: 0.0466 - sparse_categorical_accuracy: 0.9861 - scaled_adversarial_loss: 0.0925 774/1875 [===========>..................] - ETA: 8s - loss: 0.1390 - sparse_categorical_crossentropy: 0.0465 - sparse_categorical_accuracy: 0.9861 - scaled_adversarial_loss: 0.0925 781/1875 [===========>..................] - ETA: 7s - loss: 0.1389 - sparse_categorical_crossentropy: 0.0465 - sparse_categorical_accuracy: 0.9861 - scaled_adversarial_loss: 0.0924 788/1875 [===========>..................] - ETA: 7s - loss: 0.1388 - sparse_categorical_crossentropy: 0.0465 - sparse_categorical_accuracy: 0.9861 - scaled_adversarial_loss: 0.0923 796/1875 [===========>..................] - ETA: 7s - loss: 0.1387 - sparse_categorical_crossentropy: 0.0465 - sparse_categorical_accuracy: 0.9861 - scaled_adversarial_loss: 0.0923 803/1875 [===========>..................] - ETA: 7s - loss: 0.1387 - sparse_categorical_crossentropy: 0.0464 - sparse_categorical_accuracy: 0.9861 - scaled_adversarial_loss: 0.0922 811/1875 [===========>..................] - ETA: 7s - loss: 0.1386 - sparse_categorical_crossentropy: 0.0464 - sparse_categorical_accuracy: 0.9861 - scaled_adversarial_loss: 0.0922 818/1875 [============>.................] - ETA: 7s - loss: 0.1385 - sparse_categorical_crossentropy: 0.0464 - sparse_categorical_accuracy: 0.9861 - scaled_adversarial_loss: 0.0921 825/1875 [============>.................] - ETA: 7s - loss: 0.1384 - sparse_categorical_crossentropy: 0.0464 - sparse_categorical_accuracy: 0.9861 - scaled_adversarial_loss: 0.0921 833/1875 [============>.................] - ETA: 7s - loss: 0.1384 - sparse_categorical_crossentropy: 0.0463 - sparse_categorical_accuracy: 0.9861 - scaled_adversarial_loss: 0.0920 840/1875 [============>.................] - ETA: 7s - loss: 0.1383 - sparse_categorical_crossentropy: 0.0463 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0920 848/1875 [============>.................] - ETA: 7s - loss: 0.1382 - sparse_categorical_crossentropy: 0.0463 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0919 856/1875 [============>.................] - ETA: 7s - loss: 0.1381 - sparse_categorical_crossentropy: 0.0462 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0919 863/1875 [============>.................] - ETA: 7s - loss: 0.1380 - sparse_categorical_crossentropy: 0.0462 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0918 870/1875 [============>.................] - ETA: 7s - loss: 0.1380 - sparse_categorical_crossentropy: 0.0462 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0918 877/1875 [=============>................] - ETA: 7s - loss: 0.1379 - sparse_categorical_crossentropy: 0.0462 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0917 884/1875 [=============>................] - ETA: 7s - loss: 0.1378 - sparse_categorical_crossentropy: 0.0461 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0917 891/1875 [=============>................] - ETA: 7s - loss: 0.1378 - sparse_categorical_crossentropy: 0.0461 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0916 899/1875 [=============>................] - ETA: 7s - loss: 0.1377 - sparse_categorical_crossentropy: 0.0461 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0916 907/1875 [=============>................] - ETA: 7s - loss: 0.1376 - sparse_categorical_crossentropy: 0.0461 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0915 914/1875 [=============>................] - ETA: 7s - loss: 0.1375 - sparse_categorical_crossentropy: 0.0460 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0915 921/1875 [=============>................] - ETA: 6s - loss: 0.1375 - sparse_categorical_crossentropy: 0.0460 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0915 928/1875 [=============>................] - ETA: 6s - loss: 0.1374 - sparse_categorical_crossentropy: 0.0460 - sparse_categorical_accuracy: 0.9862 - scaled_adversarial_loss: 0.0914 935/1875 [=============>................] - ETA: 6s - loss: 0.1374 - sparse_categorical_crossentropy: 0.0460 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0914 943/1875 [==============>...............] - ETA: 6s - loss: 0.1373 - sparse_categorical_crossentropy: 0.0460 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0913 951/1875 [==============>...............] - ETA: 6s - loss: 0.1372 - sparse_categorical_crossentropy: 0.0459 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0913 959/1875 [==============>...............] - ETA: 6s - loss: 0.1372 - sparse_categorical_crossentropy: 0.0459 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0912 966/1875 [==============>...............] - ETA: 6s - loss: 0.1371 - sparse_categorical_crossentropy: 0.0459 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0912 973/1875 [==============>...............] - ETA: 6s - loss: 0.1370 - sparse_categorical_crossentropy: 0.0459 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0912 980/1875 [==============>...............] - ETA: 6s - loss: 0.1370 - sparse_categorical_crossentropy: 0.0459 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0911 988/1875 [==============>...............] - ETA: 6s - loss: 0.1369 - sparse_categorical_crossentropy: 0.0458 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0911 995/1875 [==============>...............] - ETA: 6s - loss: 0.1369 - sparse_categorical_crossentropy: 0.0458 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0910 1003/1875 [===============>..............] - ETA: 6s - loss: 0.1368 - sparse_categorical_crossentropy: 0.0458 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0910 1010/1875 [===============>..............] - ETA: 6s - loss: 0.1367 - sparse_categorical_crossentropy: 0.0458 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0910 1017/1875 [===============>..............] - ETA: 6s - loss: 0.1367 - sparse_categorical_crossentropy: 0.0458 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0909 1024/1875 [===============>..............] - ETA: 6s - loss: 0.1366 - sparse_categorical_crossentropy: 0.0458 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0909 1031/1875 [===============>..............] - ETA: 6s - loss: 0.1366 - sparse_categorical_crossentropy: 0.0457 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0909 1039/1875 [===============>..............] - ETA: 6s - loss: 0.1365 - sparse_categorical_crossentropy: 0.0457 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0908 1047/1875 [===============>..............] - ETA: 6s - loss: 0.1365 - sparse_categorical_crossentropy: 0.0457 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0908 1055/1875 [===============>..............] - ETA: 5s - loss: 0.1364 - sparse_categorical_crossentropy: 0.0457 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0907 1062/1875 [===============>..............] - ETA: 5s - loss: 0.1364 - sparse_categorical_crossentropy: 0.0457 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0907 1069/1875 [================>.............] - ETA: 5s - loss: 0.1363 - sparse_categorical_crossentropy: 0.0457 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0907 1077/1875 [================>.............] - ETA: 5s - loss: 0.1363 - sparse_categorical_crossentropy: 0.0457 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0906 1084/1875 [================>.............] - ETA: 5s - loss: 0.1363 - sparse_categorical_crossentropy: 0.0456 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0906 1092/1875 [================>.............] - ETA: 5s - loss: 0.1362 - sparse_categorical_crossentropy: 0.0456 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0906 1099/1875 [================>.............] - ETA: 5s - loss: 0.1362 - sparse_categorical_crossentropy: 0.0456 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0906 1106/1875 [================>.............] - ETA: 5s - loss: 0.1361 - sparse_categorical_crossentropy: 0.0456 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0905 1113/1875 [================>.............] - ETA: 5s - loss: 0.1361 - sparse_categorical_crossentropy: 0.0456 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0905 1120/1875 [================>.............] - ETA: 5s - loss: 0.1360 - sparse_categorical_crossentropy: 0.0456 - sparse_categorical_accuracy: 0.9863 - scaled_adversarial_loss: 0.0905 1127/1875 [=================>............] - ETA: 5s - loss: 0.1360 - sparse_categorical_crossentropy: 0.0455 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0904 1134/1875 [=================>............] - ETA: 5s - loss: 0.1359 - sparse_categorical_crossentropy: 0.0455 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0904 1141/1875 [=================>............] - ETA: 5s - loss: 0.1359 - sparse_categorical_crossentropy: 0.0455 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0904 1149/1875 [=================>............] - ETA: 5s - loss: 0.1358 - sparse_categorical_crossentropy: 0.0455 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0903 1156/1875 [=================>............] - ETA: 5s - loss: 0.1358 - sparse_categorical_crossentropy: 0.0455 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0903 1163/1875 [=================>............] - ETA: 5s - loss: 0.1357 - sparse_categorical_crossentropy: 0.0455 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0903 1170/1875 [=================>............] - ETA: 5s - loss: 0.1357 - sparse_categorical_crossentropy: 0.0454 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0902 1178/1875 [=================>............] - ETA: 5s - loss: 0.1356 - sparse_categorical_crossentropy: 0.0454 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0902 1185/1875 [=================>............] - ETA: 5s - loss: 0.1356 - sparse_categorical_crossentropy: 0.0454 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0901 1193/1875 [==================>...........] - ETA: 4s - loss: 0.1355 - sparse_categorical_crossentropy: 0.0454 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0901 1201/1875 [==================>...........] - ETA: 4s - loss: 0.1354 - sparse_categorical_crossentropy: 0.0454 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0901 1208/1875 [==================>...........] - ETA: 4s - loss: 0.1354 - sparse_categorical_crossentropy: 0.0454 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0900 1215/1875 [==================>...........] - ETA: 4s - loss: 0.1353 - sparse_categorical_crossentropy: 0.0453 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0900 1222/1875 [==================>...........] - ETA: 4s - loss: 0.1353 - sparse_categorical_crossentropy: 0.0453 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0900 1229/1875 [==================>...........] - ETA: 4s - loss: 0.1352 - sparse_categorical_crossentropy: 0.0453 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0899 1236/1875 [==================>...........] - ETA: 4s - loss: 0.1352 - sparse_categorical_crossentropy: 0.0453 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0899 1243/1875 [==================>...........] - ETA: 4s - loss: 0.1351 - sparse_categorical_crossentropy: 0.0453 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0899 1250/1875 [===================>..........] - ETA: 4s - loss: 0.1351 - sparse_categorical_crossentropy: 0.0453 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0898 1257/1875 [===================>..........] - ETA: 4s - loss: 0.1351 - sparse_categorical_crossentropy: 0.0453 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0898 1264/1875 [===================>..........] - ETA: 4s - loss: 0.1350 - sparse_categorical_crossentropy: 0.0452 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0898 1272/1875 [===================>..........] - ETA: 4s - loss: 0.1349 - sparse_categorical_crossentropy: 0.0452 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0897 1281/1875 [===================>..........] - ETA: 4s - loss: 0.1349 - sparse_categorical_crossentropy: 0.0452 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0897 1289/1875 [===================>..........] - ETA: 4s - loss: 0.1348 - sparse_categorical_crossentropy: 0.0452 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0896 1297/1875 [===================>..........] - ETA: 4s - loss: 0.1348 - sparse_categorical_crossentropy: 0.0452 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0896 1305/1875 [===================>..........] - ETA: 4s - loss: 0.1347 - sparse_categorical_crossentropy: 0.0451 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0896 1312/1875 [===================>..........] - ETA: 4s - loss: 0.1346 - sparse_categorical_crossentropy: 0.0451 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0895 1320/1875 [====================>.........] - ETA: 4s - loss: 0.1346 - sparse_categorical_crossentropy: 0.0451 - sparse_categorical_accuracy: 0.9864 - scaled_adversarial_loss: 0.0895 1328/1875 [====================>.........] - ETA: 3s - loss: 0.1345 - sparse_categorical_crossentropy: 0.0451 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0894 1336/1875 [====================>.........] - ETA: 3s - loss: 0.1345 - sparse_categorical_crossentropy: 0.0451 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0894 1344/1875 [====================>.........] - ETA: 3s - loss: 0.1344 - sparse_categorical_crossentropy: 0.0450 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0894 1352/1875 [====================>.........] - ETA: 3s - loss: 0.1344 - sparse_categorical_crossentropy: 0.0450 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0893 1360/1875 [====================>.........] - ETA: 3s - loss: 0.1343 - sparse_categorical_crossentropy: 0.0450 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0893 1368/1875 [====================>.........] - ETA: 3s - loss: 0.1342 - sparse_categorical_crossentropy: 0.0450 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0892 1376/1875 [=====================>........] - ETA: 3s - loss: 0.1342 - sparse_categorical_crossentropy: 0.0450 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0892 1384/1875 [=====================>........] - ETA: 3s - loss: 0.1341 - sparse_categorical_crossentropy: 0.0449 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0892 1392/1875 [=====================>........] - ETA: 3s - loss: 0.1341 - sparse_categorical_crossentropy: 0.0449 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0891 1400/1875 [=====================>........] - ETA: 3s - loss: 0.1340 - sparse_categorical_crossentropy: 0.0449 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0891 1408/1875 [=====================>........] - ETA: 3s - loss: 0.1339 - sparse_categorical_crossentropy: 0.0449 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0891 1416/1875 [=====================>........] - ETA: 3s - loss: 0.1339 - sparse_categorical_crossentropy: 0.0449 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0890 1424/1875 [=====================>........] - ETA: 3s - loss: 0.1338 - sparse_categorical_crossentropy: 0.0448 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0890 1432/1875 [=====================>........] - ETA: 3s - loss: 0.1338 - sparse_categorical_crossentropy: 0.0448 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0889 1440/1875 [======================>.......] - ETA: 3s - loss: 0.1337 - sparse_categorical_crossentropy: 0.0448 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0889 1449/1875 [======================>.......] - ETA: 3s - loss: 0.1336 - sparse_categorical_crossentropy: 0.0448 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0889 1457/1875 [======================>.......] - ETA: 3s - loss: 0.1336 - sparse_categorical_crossentropy: 0.0448 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0888 1465/1875 [======================>.......] - ETA: 2s - loss: 0.1335 - sparse_categorical_crossentropy: 0.0447 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0888 1473/1875 [======================>.......] - ETA: 2s - loss: 0.1335 - sparse_categorical_crossentropy: 0.0447 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0887 1481/1875 [======================>.......] - ETA: 2s - loss: 0.1334 - sparse_categorical_crossentropy: 0.0447 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0887 1489/1875 [======================>.......] - ETA: 2s - loss: 0.1334 - sparse_categorical_crossentropy: 0.0447 - sparse_categorical_accuracy: 0.9865 - scaled_adversarial_loss: 0.0887 1497/1875 [======================>.......] - ETA: 2s - loss: 0.1333 - sparse_categorical_crossentropy: 0.0447 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0886 1506/1875 [=======================>......] - ETA: 2s - loss: 0.1332 - sparse_categorical_crossentropy: 0.0446 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0886 1514/1875 [=======================>......] - ETA: 2s - loss: 0.1332 - sparse_categorical_crossentropy: 0.0446 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0885 1522/1875 [=======================>......] - ETA: 2s - loss: 0.1331 - sparse_categorical_crossentropy: 0.0446 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0885 1530/1875 [=======================>......] - ETA: 2s - loss: 0.1331 - sparse_categorical_crossentropy: 0.0446 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0885 1538/1875 [=======================>......] - ETA: 2s - loss: 0.1330 - sparse_categorical_crossentropy: 0.0446 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0884 1546/1875 [=======================>......] - ETA: 2s - loss: 0.1329 - sparse_categorical_crossentropy: 0.0445 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0884 1555/1875 [=======================>......] - ETA: 2s - loss: 0.1329 - sparse_categorical_crossentropy: 0.0445 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0884 1563/1875 [========================>.....] - ETA: 2s - loss: 0.1328 - sparse_categorical_crossentropy: 0.0445 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0883 1572/1875 [========================>.....] - ETA: 2s - loss: 0.1328 - sparse_categorical_crossentropy: 0.0445 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0883 1582/1875 [========================>.....] - ETA: 2s - loss: 0.1327 - sparse_categorical_crossentropy: 0.0445 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0882 1592/1875 [========================>.....] - ETA: 2s - loss: 0.1326 - sparse_categorical_crossentropy: 0.0444 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0882 1601/1875 [========================>.....] - ETA: 1s - loss: 0.1326 - sparse_categorical_crossentropy: 0.0444 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0881 1611/1875 [========================>.....] - ETA: 1s - loss: 0.1325 - sparse_categorical_crossentropy: 0.0444 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0881 1621/1875 [========================>.....] - ETA: 1s - loss: 0.1324 - sparse_categorical_crossentropy: 0.0444 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0881 1631/1875 [=========================>....] - ETA: 1s - loss: 0.1324 - sparse_categorical_crossentropy: 0.0443 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0880 1641/1875 [=========================>....] - ETA: 1s - loss: 0.1323 - sparse_categorical_crossentropy: 0.0443 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0880 1650/1875 [=========================>....] - ETA: 1s - loss: 0.1322 - sparse_categorical_crossentropy: 0.0443 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0879 1660/1875 [=========================>....] - ETA: 1s - loss: 0.1322 - sparse_categorical_crossentropy: 0.0443 - sparse_categorical_accuracy: 0.9866 - scaled_adversarial_loss: 0.0879 1670/1875 [=========================>....] - ETA: 1s - loss: 0.1321 - sparse_categorical_crossentropy: 0.0443 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0878 1680/1875 [=========================>....] - ETA: 1s - loss: 0.1320 - sparse_categorical_crossentropy: 0.0442 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0878 1690/1875 [==========================>...] - ETA: 1s - loss: 0.1320 - sparse_categorical_crossentropy: 0.0442 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0878 1701/1875 [==========================>...] - ETA: 1s - loss: 0.1319 - sparse_categorical_crossentropy: 0.0442 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0877 1711/1875 [==========================>...] - ETA: 1s - loss: 0.1319 - sparse_categorical_crossentropy: 0.0442 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0877 1721/1875 [==========================>...] - ETA: 1s - loss: 0.1318 - sparse_categorical_crossentropy: 0.0442 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0876 1731/1875 [==========================>...] - ETA: 1s - loss: 0.1317 - sparse_categorical_crossentropy: 0.0441 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0876 1742/1875 [==========================>...] - ETA: 0s - loss: 0.1317 - sparse_categorical_crossentropy: 0.0441 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0876 1753/1875 [===========================>..] - ETA: 0s - loss: 0.1316 - sparse_categorical_crossentropy: 0.0441 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0875 1763/1875 [===========================>..] - ETA: 0s - loss: 0.1315 - sparse_categorical_crossentropy: 0.0441 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0875 1773/1875 [===========================>..] - ETA: 0s - loss: 0.1315 - sparse_categorical_crossentropy: 0.0440 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0874 1783/1875 [===========================>..] - ETA: 0s - loss: 0.1314 - sparse_categorical_crossentropy: 0.0440 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0874 1793/1875 [===========================>..] - ETA: 0s - loss: 0.1313 - sparse_categorical_crossentropy: 0.0440 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0873 1804/1875 [===========================>..] - ETA: 0s - loss: 0.1313 - sparse_categorical_crossentropy: 0.0440 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0873 1814/1875 [============================>.] - ETA: 0s - loss: 0.1312 - sparse_categorical_crossentropy: 0.0440 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0873 1824/1875 [============================>.] - ETA: 0s - loss: 0.1312 - sparse_categorical_crossentropy: 0.0439 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0872 1834/1875 [============================>.] - ETA: 0s - loss: 0.1311 - sparse_categorical_crossentropy: 0.0439 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0872 1843/1875 [============================>.] - ETA: 0s - loss: 0.1310 - sparse_categorical_crossentropy: 0.0439 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0871 1853/1875 [============================>.] - ETA: 0s - loss: 0.1310 - sparse_categorical_crossentropy: 0.0439 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0871 1863/1875 [============================>.] - ETA: 0s - loss: 0.1309 - sparse_categorical_crossentropy: 0.0439 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0871 1873/1875 [============================>.] - ETA: 0s - loss: 0.1309 - sparse_categorical_crossentropy: 0.0438 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0870 1875/1875 [==============================] - 13s 7ms/step - loss: 0.1308 - sparse_categorical_crossentropy: 0.0438 - sparse_categorical_accuracy: 0.9867 - scaled_adversarial_loss: 0.0870 Epoch 3/5 1/1875 [..............................] - ETA: 12:16 - loss: 0.1308 - sparse_categorical_crossentropy: 0.0635 - sparse_categorical_accuracy: 0.9688 - scaled_adversarial_loss: 0.0673 9/1875 [..............................] - ETA: 12s - loss: 0.0987 - sparse_categorical_crossentropy: 0.0335 - sparse_categorical_accuracy: 0.9879 - scaled_adversarial_loss: 0.0652 16/1875 [..............................] - ETA: 13s - loss: 0.1029 - sparse_categorical_crossentropy: 0.0320 - sparse_categorical_accuracy: 0.9893 - scaled_adversarial_loss: 0.0709 23/1875 [..............................] - ETA: 13s - loss: 0.1016 - sparse_categorical_crossentropy: 0.0303 - sparse_categorical_accuracy: 0.9902 - scaled_adversarial_loss: 0.0714 31/1875 [..............................] - ETA: 12s - loss: 0.1006 - sparse_categorical_crossentropy: 0.0292 - sparse_categorical_accuracy: 0.9906 - scaled_adversarial_loss: 0.0715 39/1875 [..............................] - ETA: 12s - loss: 0.0985 - sparse_categorical_crossentropy: 0.0280 - sparse_categorical_accuracy: 0.9910 - scaled_adversarial_loss: 0.0705 47/1875 [..............................] - ETA: 12s - loss: 0.0975 - sparse_categorical_crossentropy: 0.0275 - sparse_categorical_accuracy: 0.9911 - scaled_adversarial_loss: 0.0701 55/1875 [..............................] - ETA: 12s - loss: 0.0994 - sparse_categorical_crossentropy: 0.0289 - sparse_categorical_accuracy: 0.9909 - scaled_adversarial_loss: 0.0704 63/1875 [>.............................] - ETA: 12s - loss: 0.1009 - sparse_categorical_crossentropy: 0.0302 - sparse_categorical_accuracy: 0.9907 - scaled_adversarial_loss: 0.0707 71/1875 [>.............................] - ETA: 12s - loss: 0.1018 - sparse_categorical_crossentropy: 0.0311 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0707 79/1875 [>.............................] - ETA: 12s - loss: 0.1024 - sparse_categorical_crossentropy: 0.0318 - sparse_categorical_accuracy: 0.9903 - scaled_adversarial_loss: 0.0706 88/1875 [>.............................] - ETA: 12s - loss: 0.1025 - sparse_categorical_crossentropy: 0.0322 - sparse_categorical_accuracy: 0.9902 - scaled_adversarial_loss: 0.0703 96/1875 [>.............................] - ETA: 12s - loss: 0.1026 - sparse_categorical_crossentropy: 0.0325 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0701 105/1875 [>.............................] - ETA: 11s - loss: 0.1024 - sparse_categorical_crossentropy: 0.0326 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0698 113/1875 [>.............................] - ETA: 11s - loss: 0.1021 - sparse_categorical_crossentropy: 0.0327 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0694 121/1875 [>.............................] - ETA: 11s - loss: 0.1017 - sparse_categorical_crossentropy: 0.0326 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0691 129/1875 [=>............................] - ETA: 11s - loss: 0.1012 - sparse_categorical_crossentropy: 0.0325 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0687 138/1875 [=>............................] - ETA: 11s - loss: 0.1008 - sparse_categorical_crossentropy: 0.0325 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0684 147/1875 [=>............................] - ETA: 11s - loss: 0.1005 - sparse_categorical_crossentropy: 0.0324 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0681 155/1875 [=>............................] - ETA: 11s - loss: 0.1002 - sparse_categorical_crossentropy: 0.0324 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0679 163/1875 [=>............................] - ETA: 11s - loss: 0.1000 - sparse_categorical_crossentropy: 0.0323 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0677 170/1875 [=>............................] - ETA: 11s - loss: 0.0998 - sparse_categorical_crossentropy: 0.0323 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0675 178/1875 [=>............................] - ETA: 11s - loss: 0.0997 - sparse_categorical_crossentropy: 0.0323 - sparse_categorical_accuracy: 0.9899 - scaled_adversarial_loss: 0.0674 186/1875 [=>............................] - ETA: 11s - loss: 0.0996 - sparse_categorical_crossentropy: 0.0323 - sparse_categorical_accuracy: 0.9899 - scaled_adversarial_loss: 0.0673 195/1875 [==>...........................] - ETA: 11s - loss: 0.0994 - sparse_categorical_crossentropy: 0.0322 - sparse_categorical_accuracy: 0.9899 - scaled_adversarial_loss: 0.0671 203/1875 [==>...........................] - ETA: 11s - loss: 0.0992 - sparse_categorical_crossentropy: 0.0322 - sparse_categorical_accuracy: 0.9899 - scaled_adversarial_loss: 0.0670 211/1875 [==>...........................] - ETA: 11s - loss: 0.0990 - sparse_categorical_crossentropy: 0.0321 - sparse_categorical_accuracy: 0.9899 - scaled_adversarial_loss: 0.0669 219/1875 [==>...........................] - ETA: 11s - loss: 0.0988 - sparse_categorical_crossentropy: 0.0321 - sparse_categorical_accuracy: 0.9899 - scaled_adversarial_loss: 0.0668 227/1875 [==>...........................] - ETA: 11s - loss: 0.0986 - sparse_categorical_crossentropy: 0.0320 - sparse_categorical_accuracy: 0.9899 - scaled_adversarial_loss: 0.0666 235/1875 [==>...........................] - ETA: 10s - loss: 0.0985 - sparse_categorical_crossentropy: 0.0320 - sparse_categorical_accuracy: 0.9899 - scaled_adversarial_loss: 0.0665 243/1875 [==>...........................] - ETA: 10s - loss: 0.0983 - sparse_categorical_crossentropy: 0.0319 - sparse_categorical_accuracy: 0.9899 - scaled_adversarial_loss: 0.0664 252/1875 [===>..........................] - ETA: 10s - loss: 0.0981 - sparse_categorical_crossentropy: 0.0318 - sparse_categorical_accuracy: 0.9899 - scaled_adversarial_loss: 0.0662 260/1875 [===>..........................] - ETA: 10s - loss: 0.0979 - sparse_categorical_crossentropy: 0.0318 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0661 268/1875 [===>..........................] - ETA: 10s - loss: 0.0977 - sparse_categorical_crossentropy: 0.0317 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0660 276/1875 [===>..........................] - ETA: 10s - loss: 0.0976 - sparse_categorical_crossentropy: 0.0316 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0659 284/1875 [===>..........................] - ETA: 10s - loss: 0.0974 - sparse_categorical_crossentropy: 0.0316 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0658 292/1875 [===>..........................] - ETA: 10s - loss: 0.0972 - sparse_categorical_crossentropy: 0.0315 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0657 300/1875 [===>..........................] - ETA: 10s - loss: 0.0970 - sparse_categorical_crossentropy: 0.0314 - sparse_categorical_accuracy: 0.9900 - scaled_adversarial_loss: 0.0656 308/1875 [===>..........................] - ETA: 10s - loss: 0.0968 - sparse_categorical_crossentropy: 0.0313 - sparse_categorical_accuracy: 0.9901 - scaled_adversarial_loss: 0.0654 316/1875 [====>.........................] - ETA: 10s - loss: 0.0966 - sparse_categorical_crossentropy: 0.0312 - sparse_categorical_accuracy: 0.9901 - scaled_adversarial_loss: 0.0653 324/1875 [====>.........................] - ETA: 10s - loss: 0.0963 - sparse_categorical_crossentropy: 0.0311 - sparse_categorical_accuracy: 0.9901 - scaled_adversarial_loss: 0.0652 332/1875 [====>.........................] - ETA: 10s - loss: 0.0961 - sparse_categorical_crossentropy: 0.0311 - sparse_categorical_accuracy: 0.9901 - scaled_adversarial_loss: 0.0651 340/1875 [====>.........................] - ETA: 10s - loss: 0.0959 - sparse_categorical_crossentropy: 0.0310 - sparse_categorical_accuracy: 0.9902 - scaled_adversarial_loss: 0.0649 348/1875 [====>.........................] - ETA: 10s - loss: 0.0957 - sparse_categorical_crossentropy: 0.0309 - sparse_categorical_accuracy: 0.9902 - scaled_adversarial_loss: 0.0648 356/1875 [====>.........................] - ETA: 10s - loss: 0.0956 - sparse_categorical_crossentropy: 0.0308 - sparse_categorical_accuracy: 0.9902 - scaled_adversarial_loss: 0.0647 364/1875 [====>.........................] - ETA: 10s - loss: 0.0954 - sparse_categorical_crossentropy: 0.0307 - sparse_categorical_accuracy: 0.9902 - scaled_adversarial_loss: 0.0646 372/1875 [====>.........................] - ETA: 9s - loss: 0.0952 - sparse_categorical_crossentropy: 0.0307 - sparse_categorical_accuracy: 0.9902 - scaled_adversarial_loss: 0.0645 380/1875 [=====>........................] - ETA: 9s - loss: 0.0950 - sparse_categorical_crossentropy: 0.0306 - sparse_categorical_accuracy: 0.9903 - scaled_adversarial_loss: 0.0644 388/1875 [=====>........................] - ETA: 9s - loss: 0.0949 - sparse_categorical_crossentropy: 0.0306 - sparse_categorical_accuracy: 0.9903 - scaled_adversarial_loss: 0.0643 396/1875 [=====>........................] - ETA: 9s - loss: 0.0947 - sparse_categorical_crossentropy: 0.0305 - sparse_categorical_accuracy: 0.9903 - scaled_adversarial_loss: 0.0642 405/1875 [=====>........................] - ETA: 9s - loss: 0.0946 - sparse_categorical_crossentropy: 0.0304 - sparse_categorical_accuracy: 0.9903 - scaled_adversarial_loss: 0.0642 414/1875 [=====>........................] - ETA: 9s - loss: 0.0945 - sparse_categorical_crossentropy: 0.0304 - sparse_categorical_accuracy: 0.9903 - scaled_adversarial_loss: 0.0641 423/1875 [=====>........................] - ETA: 9s - loss: 0.0944 - sparse_categorical_crossentropy: 0.0304 - sparse_categorical_accuracy: 0.9903 - scaled_adversarial_loss: 0.0640 431/1875 [=====>........................] - ETA: 9s - loss: 0.0943 - sparse_categorical_crossentropy: 0.0303 - sparse_categorical_accuracy: 0.9903 - scaled_adversarial_loss: 0.0640 439/1875 [======>.......................] - ETA: 9s - loss: 0.0943 - sparse_categorical_crossentropy: 0.0303 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0640 447/1875 [======>.......................] - ETA: 9s - loss: 0.0942 - sparse_categorical_crossentropy: 0.0303 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0639 456/1875 [======>.......................] - ETA: 9s - loss: 0.0941 - sparse_categorical_crossentropy: 0.0303 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0639 464/1875 [======>.......................] - ETA: 9s - loss: 0.0940 - sparse_categorical_crossentropy: 0.0302 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0638 472/1875 [======>.......................] - ETA: 9s - loss: 0.0940 - sparse_categorical_crossentropy: 0.0302 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0638 480/1875 [======>.......................] - ETA: 9s - loss: 0.0939 - sparse_categorical_crossentropy: 0.0302 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0637 488/1875 [======>.......................] - ETA: 9s - loss: 0.0939 - sparse_categorical_crossentropy: 0.0302 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0637 495/1875 [======>.......................] - ETA: 9s - loss: 0.0938 - sparse_categorical_crossentropy: 0.0301 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0637 503/1875 [=======>......................] - ETA: 9s - loss: 0.0937 - sparse_categorical_crossentropy: 0.0301 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0636 512/1875 [=======>......................] - ETA: 8s - loss: 0.0937 - sparse_categorical_crossentropy: 0.0301 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0636 520/1875 [=======>......................] - ETA: 8s - loss: 0.0937 - sparse_categorical_crossentropy: 0.0301 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0636 528/1875 [=======>......................] - ETA: 8s - loss: 0.0936 - sparse_categorical_crossentropy: 0.0301 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0635 536/1875 [=======>......................] - ETA: 8s - loss: 0.0936 - sparse_categorical_crossentropy: 0.0300 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0635 544/1875 [=======>......................] - ETA: 8s - loss: 0.0935 - sparse_categorical_crossentropy: 0.0300 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0635 551/1875 [=======>......................] - ETA: 8s - loss: 0.0935 - sparse_categorical_crossentropy: 0.0300 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0634 559/1875 [=======>......................] - ETA: 8s - loss: 0.0934 - sparse_categorical_crossentropy: 0.0300 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0634 567/1875 [========>.....................] - ETA: 8s - loss: 0.0934 - sparse_categorical_crossentropy: 0.0300 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0634 575/1875 [========>.....................] - ETA: 8s - loss: 0.0934 - sparse_categorical_crossentropy: 0.0300 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0634 583/1875 [========>.....................] - ETA: 8s - loss: 0.0933 - sparse_categorical_crossentropy: 0.0300 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0633 591/1875 [========>.....................] - ETA: 8s - loss: 0.0933 - sparse_categorical_crossentropy: 0.0300 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0633 599/1875 [========>.....................] - ETA: 8s - loss: 0.0932 - sparse_categorical_crossentropy: 0.0300 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0633 607/1875 [========>.....................] - ETA: 8s - loss: 0.0932 - sparse_categorical_crossentropy: 0.0299 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0633 616/1875 [========>.....................] - ETA: 8s - loss: 0.0932 - sparse_categorical_crossentropy: 0.0299 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0633 625/1875 [=========>....................] - ETA: 8s - loss: 0.0932 - sparse_categorical_crossentropy: 0.0299 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0632 633/1875 [=========>....................] - ETA: 8s - loss: 0.0931 - sparse_categorical_crossentropy: 0.0299 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0632 641/1875 [=========>....................] - ETA: 8s - loss: 0.0931 - sparse_categorical_crossentropy: 0.0299 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0632 648/1875 [=========>....................] - ETA: 8s - loss: 0.0931 - sparse_categorical_crossentropy: 0.0299 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0632 656/1875 [=========>....................] - ETA: 8s - loss: 0.0930 - sparse_categorical_crossentropy: 0.0299 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0632 665/1875 [=========>....................] - ETA: 7s - loss: 0.0930 - sparse_categorical_crossentropy: 0.0298 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0632 674/1875 [=========>....................] - ETA: 7s - loss: 0.0930 - sparse_categorical_crossentropy: 0.0298 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0631 682/1875 [=========>....................] - ETA: 7s - loss: 0.0929 - sparse_categorical_crossentropy: 0.0298 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0631 690/1875 [==========>...................] - ETA: 7s - loss: 0.0929 - sparse_categorical_crossentropy: 0.0298 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0631 698/1875 [==========>...................] - ETA: 7s - loss: 0.0928 - sparse_categorical_crossentropy: 0.0298 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0631 706/1875 [==========>...................] - ETA: 7s - loss: 0.0928 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0631 714/1875 [==========>...................] - ETA: 7s - loss: 0.0928 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0630 722/1875 [==========>...................] - ETA: 7s - loss: 0.0927 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0630 730/1875 [==========>...................] - ETA: 7s - loss: 0.0927 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0630 738/1875 [==========>...................] - ETA: 7s - loss: 0.0927 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0630 746/1875 [==========>...................] - ETA: 7s - loss: 0.0926 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0630 754/1875 [===========>..................] - ETA: 7s - loss: 0.0926 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0629 762/1875 [===========>..................] - ETA: 7s - loss: 0.0925 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0629 771/1875 [===========>..................] - ETA: 7s - loss: 0.0925 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0629 779/1875 [===========>..................] - ETA: 7s - loss: 0.0925 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0629 788/1875 [===========>..................] - ETA: 7s - loss: 0.0924 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0628 796/1875 [===========>..................] - ETA: 7s - loss: 0.0924 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0628 803/1875 [===========>..................] - ETA: 7s - loss: 0.0924 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0628 810/1875 [===========>..................] - ETA: 7s - loss: 0.0923 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0628 818/1875 [============>.................] - ETA: 6s - loss: 0.0923 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0628 826/1875 [============>.................] - ETA: 6s - loss: 0.0923 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0627 835/1875 [============>.................] - ETA: 6s - loss: 0.0923 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0627 843/1875 [============>.................] - ETA: 6s - loss: 0.0923 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0627 851/1875 [============>.................] - ETA: 6s - loss: 0.0923 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0627 858/1875 [============>.................] - ETA: 6s - loss: 0.0923 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0627 866/1875 [============>.................] - ETA: 6s - loss: 0.0922 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0627 874/1875 [============>.................] - ETA: 6s - loss: 0.0922 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0627 882/1875 [=============>................] - ETA: 6s - loss: 0.0922 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0627 891/1875 [=============>................] - ETA: 6s - loss: 0.0922 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0626 900/1875 [=============>................] - ETA: 6s - loss: 0.0922 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0626 908/1875 [=============>................] - ETA: 6s - loss: 0.0922 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0626 916/1875 [=============>................] - ETA: 6s - loss: 0.0922 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0626 924/1875 [=============>................] - ETA: 6s - loss: 0.0922 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0626 932/1875 [=============>................] - ETA: 6s - loss: 0.0922 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0626 940/1875 [==============>...............] - ETA: 6s - loss: 0.0921 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0625 948/1875 [==============>...............] - ETA: 6s - loss: 0.0921 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0625 956/1875 [==============>...............] - ETA: 6s - loss: 0.0921 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0625 964/1875 [==============>...............] - ETA: 5s - loss: 0.0921 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0625 972/1875 [==============>...............] - ETA: 5s - loss: 0.0921 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0625 980/1875 [==============>...............] - ETA: 5s - loss: 0.0921 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0625 988/1875 [==============>...............] - ETA: 5s - loss: 0.0921 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0625 996/1875 [==============>...............] - ETA: 5s - loss: 0.0920 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0624 1005/1875 [===============>..............] - ETA: 5s - loss: 0.0920 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0624 1013/1875 [===============>..............] - ETA: 5s - loss: 0.0920 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0624 1021/1875 [===============>..............] - ETA: 5s - loss: 0.0920 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0624 1029/1875 [===============>..............] - ETA: 5s - loss: 0.0920 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0624 1037/1875 [===============>..............] - ETA: 5s - loss: 0.0920 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0624 1045/1875 [===============>..............] - ETA: 5s - loss: 0.0920 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0623 1053/1875 [===============>..............] - ETA: 5s - loss: 0.0920 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0623 1062/1875 [===============>..............] - ETA: 5s - loss: 0.0920 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0623 1070/1875 [================>.............] - ETA: 5s - loss: 0.0919 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0623 1078/1875 [================>.............] - ETA: 5s - loss: 0.0919 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0623 1086/1875 [================>.............] - ETA: 5s - loss: 0.0919 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0623 1095/1875 [================>.............] - ETA: 5s - loss: 0.0919 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0622 1103/1875 [================>.............] - ETA: 5s - loss: 0.0919 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0622 1111/1875 [================>.............] - ETA: 5s - loss: 0.0919 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0622 1119/1875 [================>.............] - ETA: 4s - loss: 0.0919 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0622 1127/1875 [=================>............] - ETA: 4s - loss: 0.0919 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0622 1135/1875 [=================>............] - ETA: 4s - loss: 0.0918 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0622 1143/1875 [=================>............] - ETA: 4s - loss: 0.0918 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0622 1151/1875 [=================>............] - ETA: 4s - loss: 0.0918 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0621 1159/1875 [=================>............] - ETA: 4s - loss: 0.0918 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0621 1168/1875 [=================>............] - ETA: 4s - loss: 0.0918 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0621 1176/1875 [=================>............] - ETA: 4s - loss: 0.0918 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0621 1184/1875 [=================>............] - ETA: 4s - loss: 0.0917 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0621 1192/1875 [==================>...........] - ETA: 4s - loss: 0.0917 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0621 1201/1875 [==================>...........] - ETA: 4s - loss: 0.0917 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0620 1209/1875 [==================>...........] - ETA: 4s - loss: 0.0917 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0620 1218/1875 [==================>...........] - ETA: 4s - loss: 0.0917 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0620 1226/1875 [==================>...........] - ETA: 4s - loss: 0.0917 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0620 1234/1875 [==================>...........] - ETA: 4s - loss: 0.0917 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0620 1242/1875 [==================>...........] - ETA: 4s - loss: 0.0916 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0619 1250/1875 [===================>..........] - ETA: 4s - loss: 0.0916 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0619 1258/1875 [===================>..........] - ETA: 4s - loss: 0.0916 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0619 1266/1875 [===================>..........] - ETA: 3s - loss: 0.0916 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0619 1274/1875 [===================>..........] - ETA: 3s - loss: 0.0916 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0619 1282/1875 [===================>..........] - ETA: 3s - loss: 0.0916 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0619 1290/1875 [===================>..........] - ETA: 3s - loss: 0.0915 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0618 1298/1875 [===================>..........] - ETA: 3s - loss: 0.0915 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0618 1307/1875 [===================>..........] - ETA: 3s - loss: 0.0915 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0618 1315/1875 [====================>.........] - ETA: 3s - loss: 0.0915 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0618 1323/1875 [====================>.........] - ETA: 3s - loss: 0.0915 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0618 1331/1875 [====================>.........] - ETA: 3s - loss: 0.0914 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0617 1339/1875 [====================>.........] - ETA: 3s - loss: 0.0914 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0617 1347/1875 [====================>.........] - ETA: 3s - loss: 0.0914 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0617 1355/1875 [====================>.........] - ETA: 3s - loss: 0.0914 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0617 1364/1875 [====================>.........] - ETA: 3s - loss: 0.0914 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0617 1373/1875 [====================>.........] - ETA: 3s - loss: 0.0914 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0616 1381/1875 [=====================>........] - ETA: 3s - loss: 0.0913 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0616 1389/1875 [=====================>........] - ETA: 3s - loss: 0.0913 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0616 1397/1875 [=====================>........] - ETA: 3s - loss: 0.0913 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0616 1406/1875 [=====================>........] - ETA: 3s - loss: 0.0913 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0616 1415/1875 [=====================>........] - ETA: 3s - loss: 0.0913 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0616 1423/1875 [=====================>........] - ETA: 2s - loss: 0.0913 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0615 1432/1875 [=====================>........] - ETA: 2s - loss: 0.0912 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0615 1440/1875 [======================>.......] - ETA: 2s - loss: 0.0912 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0615 1448/1875 [======================>.......] - ETA: 2s - loss: 0.0912 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0615 1456/1875 [======================>.......] - ETA: 2s - loss: 0.0912 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0615 1465/1875 [======================>.......] - ETA: 2s - loss: 0.0912 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0615 1473/1875 [======================>.......] - ETA: 2s - loss: 0.0912 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0614 1482/1875 [======================>.......] - ETA: 2s - loss: 0.0912 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0614 1491/1875 [======================>.......] - ETA: 2s - loss: 0.0911 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0614 1499/1875 [======================>.......] - ETA: 2s - loss: 0.0911 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0614 1507/1875 [=======================>......] - ETA: 2s - loss: 0.0911 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0614 1516/1875 [=======================>......] - ETA: 2s - loss: 0.0911 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0614 1525/1875 [=======================>......] - ETA: 2s - loss: 0.0911 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0613 1533/1875 [=======================>......] - ETA: 2s - loss: 0.0911 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0613 1541/1875 [=======================>......] - ETA: 2s - loss: 0.0910 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0613 1549/1875 [=======================>......] - ETA: 2s - loss: 0.0910 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0613 1557/1875 [=======================>......] - ETA: 2s - loss: 0.0910 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0613 1564/1875 [========================>.....] - ETA: 2s - loss: 0.0910 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0612 1575/1875 [========================>.....] - ETA: 1s - loss: 0.0910 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0612 1586/1875 [========================>.....] - ETA: 1s - loss: 0.0909 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0612 1596/1875 [========================>.....] - ETA: 1s - loss: 0.0909 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0612 1606/1875 [========================>.....] - ETA: 1s - loss: 0.0909 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0612 1616/1875 [========================>.....] - ETA: 1s - loss: 0.0909 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0611 1626/1875 [=========================>....] - ETA: 1s - loss: 0.0908 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9904 - scaled_adversarial_loss: 0.0611 1637/1875 [=========================>....] - ETA: 1s - loss: 0.0908 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0611 1648/1875 [=========================>....] - ETA: 1s - loss: 0.0908 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0610 1659/1875 [=========================>....] - ETA: 1s - loss: 0.0907 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0610 1670/1875 [=========================>....] - ETA: 1s - loss: 0.0907 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0610 1680/1875 [=========================>....] - ETA: 1s - loss: 0.0907 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0610 1689/1875 [==========================>...] - ETA: 1s - loss: 0.0907 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0609 1699/1875 [==========================>...] - ETA: 1s - loss: 0.0906 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0609 1710/1875 [==========================>...] - ETA: 1s - loss: 0.0906 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0609 1720/1875 [==========================>...] - ETA: 0s - loss: 0.0906 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0609 1730/1875 [==========================>...] - ETA: 0s - loss: 0.0905 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0608 1740/1875 [==========================>...] - ETA: 0s - loss: 0.0905 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0608 1750/1875 [===========================>..] - ETA: 0s - loss: 0.0905 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0608 1760/1875 [===========================>..] - ETA: 0s - loss: 0.0904 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0608 1770/1875 [===========================>..] - ETA: 0s - loss: 0.0904 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0607 1780/1875 [===========================>..] - ETA: 0s - loss: 0.0904 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0607 1791/1875 [===========================>..] - ETA: 0s - loss: 0.0903 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0607 1802/1875 [===========================>..] - ETA: 0s - loss: 0.0903 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0606 1812/1875 [===========================>..] - ETA: 0s - loss: 0.0903 - sparse_categorical_crossentropy: 0.0297 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0606 1822/1875 [============================>.] - ETA: 0s - loss: 0.0902 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0606 1832/1875 [============================>.] - ETA: 0s - loss: 0.0902 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0606 1842/1875 [============================>.] - ETA: 0s - loss: 0.0902 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0605 1852/1875 [============================>.] - ETA: 0s - loss: 0.0901 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0605 1862/1875 [============================>.] - ETA: 0s - loss: 0.0901 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0605 1872/1875 [============================>.] - ETA: 0s - loss: 0.0901 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0604 1875/1875 [==============================] - 12s 6ms/step - loss: 0.0901 - sparse_categorical_crossentropy: 0.0296 - sparse_categorical_accuracy: 0.9905 - scaled_adversarial_loss: 0.0604 Epoch 4/5 1/1875 [..............................] - ETA: 12:43 - loss: 0.1893 - sparse_categorical_crossentropy: 0.1333 - sparse_categorical_accuracy: 0.9688 - scaled_adversarial_loss: 0.0560 9/1875 [..............................] - ETA: 12s - loss: 0.0837 - sparse_categorical_crossentropy: 0.0479 - sparse_categorical_accuracy: 0.9902 - scaled_adversarial_loss: 0.0358 17/1875 [..............................] - ETA: 12s - loss: 0.0660 - sparse_categorical_crossentropy: 0.0341 - sparse_categorical_accuracy: 0.9935 - scaled_adversarial_loss: 0.0319 25/1875 [..............................] - ETA: 12s - loss: 0.0581 - sparse_categorical_crossentropy: 0.0290 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0291 33/1875 [..............................] - ETA: 12s - loss: 0.0539 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9949 - scaled_adversarial_loss: 0.0278 42/1875 [..............................] - ETA: 12s - loss: 0.0549 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9950 - scaled_adversarial_loss: 0.0296 50/1875 [..............................] - ETA: 12s - loss: 0.0565 - sparse_categorical_crossentropy: 0.0257 - sparse_categorical_accuracy: 0.9948 - scaled_adversarial_loss: 0.0307 58/1875 [..............................] - ETA: 12s - loss: 0.0578 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0317 66/1875 [>.............................] - ETA: 11s - loss: 0.0590 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0328 74/1875 [>.............................] - ETA: 11s - loss: 0.0593 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0333 82/1875 [>.............................] - ETA: 11s - loss: 0.0595 - sparse_categorical_crossentropy: 0.0258 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0336 90/1875 [>.............................] - ETA: 11s - loss: 0.0595 - sparse_categorical_crossentropy: 0.0256 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0339 98/1875 [>.............................] - ETA: 11s - loss: 0.0595 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0341 106/1875 [>.............................] - ETA: 11s - loss: 0.0598 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0344 114/1875 [>.............................] - ETA: 11s - loss: 0.0603 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9941 - scaled_adversarial_loss: 0.0349 122/1875 [>.............................] - ETA: 11s - loss: 0.0608 - sparse_categorical_crossentropy: 0.0256 - sparse_categorical_accuracy: 0.9940 - scaled_adversarial_loss: 0.0352 130/1875 [=>............................] - ETA: 11s - loss: 0.0612 - sparse_categorical_crossentropy: 0.0257 - sparse_categorical_accuracy: 0.9938 - scaled_adversarial_loss: 0.0355 138/1875 [=>............................] - ETA: 11s - loss: 0.0615 - sparse_categorical_crossentropy: 0.0258 - sparse_categorical_accuracy: 0.9937 - scaled_adversarial_loss: 0.0358 146/1875 [=>............................] - ETA: 11s - loss: 0.0618 - sparse_categorical_crossentropy: 0.0259 - sparse_categorical_accuracy: 0.9936 - scaled_adversarial_loss: 0.0359 154/1875 [=>............................] - ETA: 11s - loss: 0.0620 - sparse_categorical_crossentropy: 0.0259 - sparse_categorical_accuracy: 0.9935 - scaled_adversarial_loss: 0.0361 162/1875 [=>............................] - ETA: 11s - loss: 0.0622 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9934 - scaled_adversarial_loss: 0.0362 170/1875 [=>............................] - ETA: 11s - loss: 0.0624 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9933 - scaled_adversarial_loss: 0.0363 178/1875 [=>............................] - ETA: 11s - loss: 0.0625 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9932 - scaled_adversarial_loss: 0.0364 186/1875 [=>............................] - ETA: 11s - loss: 0.0626 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9932 - scaled_adversarial_loss: 0.0365 194/1875 [==>...........................] - ETA: 11s - loss: 0.0627 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9931 - scaled_adversarial_loss: 0.0366 202/1875 [==>...........................] - ETA: 11s - loss: 0.0628 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9931 - scaled_adversarial_loss: 0.0367 209/1875 [==>...........................] - ETA: 11s - loss: 0.0629 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9931 - scaled_adversarial_loss: 0.0368 217/1875 [==>...........................] - ETA: 10s - loss: 0.0629 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9930 - scaled_adversarial_loss: 0.0368 225/1875 [==>...........................] - ETA: 10s - loss: 0.0630 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9930 - scaled_adversarial_loss: 0.0368 233/1875 [==>...........................] - ETA: 10s - loss: 0.0631 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9930 - scaled_adversarial_loss: 0.0369 241/1875 [==>...........................] - ETA: 10s - loss: 0.0632 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9930 - scaled_adversarial_loss: 0.0371 249/1875 [==>...........................] - ETA: 10s - loss: 0.0634 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9929 - scaled_adversarial_loss: 0.0372 257/1875 [===>..........................] - ETA: 10s - loss: 0.0636 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9929 - scaled_adversarial_loss: 0.0374 264/1875 [===>..........................] - ETA: 10s - loss: 0.0637 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9929 - scaled_adversarial_loss: 0.0375 272/1875 [===>..........................] - ETA: 10s - loss: 0.0638 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9928 - scaled_adversarial_loss: 0.0376 280/1875 [===>..........................] - ETA: 10s - loss: 0.0640 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9928 - scaled_adversarial_loss: 0.0377 288/1875 [===>..........................] - ETA: 10s - loss: 0.0642 - sparse_categorical_crossentropy: 0.0263 - sparse_categorical_accuracy: 0.9928 - scaled_adversarial_loss: 0.0379 296/1875 [===>..........................] - ETA: 10s - loss: 0.0643 - sparse_categorical_crossentropy: 0.0263 - sparse_categorical_accuracy: 0.9927 - scaled_adversarial_loss: 0.0380 304/1875 [===>..........................] - ETA: 10s - loss: 0.0644 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9927 - scaled_adversarial_loss: 0.0382 311/1875 [===>..........................] - ETA: 10s - loss: 0.0645 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9927 - scaled_adversarial_loss: 0.0383 318/1875 [====>.........................] - ETA: 10s - loss: 0.0646 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9927 - scaled_adversarial_loss: 0.0384 326/1875 [====>.........................] - ETA: 10s - loss: 0.0648 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9927 - scaled_adversarial_loss: 0.0385 334/1875 [====>.........................] - ETA: 10s - loss: 0.0649 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0387 342/1875 [====>.........................] - ETA: 10s - loss: 0.0650 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0388 350/1875 [====>.........................] - ETA: 10s - loss: 0.0651 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0389 358/1875 [====>.........................] - ETA: 10s - loss: 0.0652 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0390 366/1875 [====>.........................] - ETA: 10s - loss: 0.0653 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0391 374/1875 [====>.........................] - ETA: 9s - loss: 0.0653 - sparse_categorical_crossentropy: 0.0262 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0391 382/1875 [=====>........................] - ETA: 9s - loss: 0.0654 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0392 390/1875 [=====>........................] - ETA: 9s - loss: 0.0654 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0393 398/1875 [=====>........................] - ETA: 9s - loss: 0.0655 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0394 407/1875 [=====>........................] - ETA: 9s - loss: 0.0656 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0395 415/1875 [=====>........................] - ETA: 9s - loss: 0.0656 - sparse_categorical_crossentropy: 0.0261 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0395 423/1875 [=====>........................] - ETA: 9s - loss: 0.0657 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0396 431/1875 [=====>........................] - ETA: 9s - loss: 0.0657 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0397 439/1875 [======>.......................] - ETA: 9s - loss: 0.0658 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0397 447/1875 [======>.......................] - ETA: 9s - loss: 0.0658 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0398 455/1875 [======>.......................] - ETA: 9s - loss: 0.0659 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0399 463/1875 [======>.......................] - ETA: 9s - loss: 0.0659 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0399 471/1875 [======>.......................] - ETA: 9s - loss: 0.0660 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0400 479/1875 [======>.......................] - ETA: 9s - loss: 0.0660 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0400 487/1875 [======>.......................] - ETA: 9s - loss: 0.0661 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0400 495/1875 [======>.......................] - ETA: 9s - loss: 0.0661 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0401 503/1875 [=======>......................] - ETA: 9s - loss: 0.0661 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0401 511/1875 [=======>......................] - ETA: 9s - loss: 0.0661 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0401 519/1875 [=======>......................] - ETA: 9s - loss: 0.0662 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0401 526/1875 [=======>......................] - ETA: 8s - loss: 0.0662 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0402 534/1875 [=======>......................] - ETA: 8s - loss: 0.0662 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0402 542/1875 [=======>......................] - ETA: 8s - loss: 0.0662 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0403 550/1875 [=======>......................] - ETA: 8s - loss: 0.0663 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0403 558/1875 [=======>......................] - ETA: 8s - loss: 0.0663 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0403 566/1875 [========>.....................] - ETA: 8s - loss: 0.0663 - sparse_categorical_crossentropy: 0.0260 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0404 573/1875 [========>.....................] - ETA: 8s - loss: 0.0663 - sparse_categorical_crossentropy: 0.0259 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0404 581/1875 [========>.....................] - ETA: 8s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0259 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0404 589/1875 [========>.....................] - ETA: 8s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0259 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0405 597/1875 [========>.....................] - ETA: 8s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0259 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0405 605/1875 [========>.....................] - ETA: 8s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0259 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0405 613/1875 [========>.....................] - ETA: 8s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0259 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0405 621/1875 [========>.....................] - ETA: 8s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0259 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0406 629/1875 [=========>....................] - ETA: 8s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0258 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0406 637/1875 [=========>....................] - ETA: 8s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0258 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0406 645/1875 [=========>....................] - ETA: 8s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0258 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0406 653/1875 [=========>....................] - ETA: 8s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0258 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0407 661/1875 [=========>....................] - ETA: 8s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0258 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0407 669/1875 [=========>....................] - ETA: 7s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0257 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0407 677/1875 [=========>....................] - ETA: 7s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0257 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0407 685/1875 [=========>....................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0257 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0408 693/1875 [==========>...................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0257 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0408 702/1875 [==========>...................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0256 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0408 711/1875 [==========>...................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0256 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0408 719/1875 [==========>...................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0256 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0409 727/1875 [==========>...................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0256 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0409 735/1875 [==========>...................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0256 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0409 743/1875 [==========>...................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0256 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0409 751/1875 [===========>..................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0256 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0410 759/1875 [===========>..................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0256 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0410 767/1875 [===========>..................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0410 775/1875 [===========>..................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0410 783/1875 [===========>..................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0410 791/1875 [===========>..................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0410 799/1875 [===========>..................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0410 807/1875 [===========>..................] - ETA: 7s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0410 815/1875 [============>.................] - ETA: 7s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0410 823/1875 [============>.................] - ETA: 6s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0411 831/1875 [============>.................] - ETA: 6s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0411 838/1875 [============>.................] - ETA: 6s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0411 846/1875 [============>.................] - ETA: 6s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0411 854/1875 [============>.................] - ETA: 6s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0411 863/1875 [============>.................] - ETA: 6s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0411 871/1875 [============>.................] - ETA: 6s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0412 879/1875 [=============>................] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0412 887/1875 [=============>................] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0412 895/1875 [=============>................] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0412 903/1875 [=============>................] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0255 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0412 911/1875 [=============>................] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0412 920/1875 [=============>................] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0413 928/1875 [=============>................] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0413 936/1875 [=============>................] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0413 944/1875 [==============>...............] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0413 951/1875 [==============>...............] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0413 959/1875 [==============>...............] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0413 967/1875 [==============>...............] - ETA: 6s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0413 974/1875 [==============>...............] - ETA: 5s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0413 982/1875 [==============>...............] - ETA: 5s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0414 990/1875 [==============>...............] - ETA: 5s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0414 998/1875 [==============>...............] - ETA: 5s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0414 1006/1875 [===============>..............] - ETA: 5s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0414 1014/1875 [===============>..............] - ETA: 5s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0414 1022/1875 [===============>..............] - ETA: 5s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0415 1031/1875 [===============>..............] - ETA: 5s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0254 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0415 1039/1875 [===============>..............] - ETA: 5s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0415 1047/1875 [===============>..............] - ETA: 5s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9924 - scaled_adversarial_loss: 0.0415 1054/1875 [===============>..............] - ETA: 5s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0415 1062/1875 [===============>..............] - ETA: 5s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0415 1070/1875 [================>.............] - ETA: 5s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0416 1078/1875 [================>.............] - ETA: 5s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0416 1086/1875 [================>.............] - ETA: 5s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0416 1094/1875 [================>.............] - ETA: 5s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0416 1102/1875 [================>.............] - ETA: 5s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0416 1110/1875 [================>.............] - ETA: 5s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0417 1118/1875 [================>.............] - ETA: 5s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0417 1126/1875 [=================>............] - ETA: 4s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0417 1134/1875 [=================>............] - ETA: 4s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0417 1142/1875 [=================>............] - ETA: 4s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0417 1150/1875 [=================>............] - ETA: 4s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0417 1158/1875 [=================>............] - ETA: 4s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0418 1167/1875 [=================>............] - ETA: 4s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0418 1175/1875 [=================>............] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0418 1183/1875 [=================>............] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0418 1191/1875 [==================>...........] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0418 1199/1875 [==================>...........] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0253 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0418 1206/1875 [==================>...........] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0418 1214/1875 [==================>...........] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0418 1222/1875 [==================>...........] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0418 1230/1875 [==================>...........] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1238/1875 [==================>...........] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1246/1875 [==================>...........] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1254/1875 [===================>..........] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1262/1875 [===================>..........] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1270/1875 [===================>..........] - ETA: 4s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1278/1875 [===================>..........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1286/1875 [===================>..........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1294/1875 [===================>..........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1302/1875 [===================>..........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1310/1875 [===================>..........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1318/1875 [====================>.........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0252 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1326/1875 [====================>.........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0251 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1334/1875 [====================>.........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0251 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1342/1875 [====================>.........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0251 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1350/1875 [====================>.........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0251 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1358/1875 [====================>.........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0251 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1366/1875 [====================>.........] - ETA: 3s - loss: 0.0671 - sparse_categorical_crossentropy: 0.0251 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1374/1875 [====================>.........] - ETA: 3s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0251 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1382/1875 [=====================>........] - ETA: 3s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0251 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1390/1875 [=====================>........] - ETA: 3s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0251 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1398/1875 [=====================>........] - ETA: 3s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0251 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1406/1875 [=====================>........] - ETA: 3s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0250 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1414/1875 [=====================>........] - ETA: 3s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0250 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1421/1875 [=====================>........] - ETA: 3s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0250 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1429/1875 [=====================>........] - ETA: 2s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0250 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1437/1875 [=====================>........] - ETA: 2s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0250 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1445/1875 [======================>.......] - ETA: 2s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0250 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1454/1875 [======================>.......] - ETA: 2s - loss: 0.0670 - sparse_categorical_crossentropy: 0.0250 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1462/1875 [======================>.......] - ETA: 2s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0250 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1469/1875 [======================>.......] - ETA: 2s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0250 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1477/1875 [======================>.......] - ETA: 2s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0250 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0420 1486/1875 [======================>.......] - ETA: 2s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0249 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0419 1495/1875 [======================>.......] - ETA: 2s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0249 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1503/1875 [=======================>......] - ETA: 2s - loss: 0.0669 - sparse_categorical_crossentropy: 0.0249 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1511/1875 [=======================>......] - ETA: 2s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0249 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1519/1875 [=======================>......] - ETA: 2s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0249 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1527/1875 [=======================>......] - ETA: 2s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0249 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1535/1875 [=======================>......] - ETA: 2s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0249 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1544/1875 [=======================>......] - ETA: 2s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0249 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1552/1875 [=======================>......] - ETA: 2s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0249 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1561/1875 [=======================>......] - ETA: 2s - loss: 0.0668 - sparse_categorical_crossentropy: 0.0248 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1569/1875 [========================>.....] - ETA: 2s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0248 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1579/1875 [========================>.....] - ETA: 1s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0248 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1590/1875 [========================>.....] - ETA: 1s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0248 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1600/1875 [========================>.....] - ETA: 1s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0248 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1611/1875 [========================>.....] - ETA: 1s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0248 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1622/1875 [========================>.....] - ETA: 1s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0248 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1633/1875 [=========================>....] - ETA: 1s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0248 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1643/1875 [=========================>....] - ETA: 1s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0248 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1653/1875 [=========================>....] - ETA: 1s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0248 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1664/1875 [=========================>....] - ETA: 1s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0248 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1675/1875 [=========================>....] - ETA: 1s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0247 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1686/1875 [=========================>....] - ETA: 1s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0247 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1696/1875 [==========================>...] - ETA: 1s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0247 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1706/1875 [==========================>...] - ETA: 1s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0247 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1716/1875 [==========================>...] - ETA: 1s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0247 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1727/1875 [==========================>...] - ETA: 0s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0247 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1738/1875 [==========================>...] - ETA: 0s - loss: 0.0666 - sparse_categorical_crossentropy: 0.0247 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1749/1875 [==========================>...] - ETA: 0s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0247 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1760/1875 [===========================>..] - ETA: 0s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0247 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1771/1875 [===========================>..] - ETA: 0s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0247 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0419 1781/1875 [===========================>..] - ETA: 0s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0246 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0418 1791/1875 [===========================>..] - ETA: 0s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0246 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0418 1802/1875 [===========================>..] - ETA: 0s - loss: 0.0665 - sparse_categorical_crossentropy: 0.0246 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0418 1812/1875 [===========================>..] - ETA: 0s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0246 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0418 1823/1875 [============================>.] - ETA: 0s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0246 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0418 1834/1875 [============================>.] - ETA: 0s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0246 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0418 1843/1875 [============================>.] - ETA: 0s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0246 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0418 1853/1875 [============================>.] - ETA: 0s - loss: 0.0664 - sparse_categorical_crossentropy: 0.0246 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0418 1864/1875 [============================>.] - ETA: 0s - loss: 0.0663 - sparse_categorical_crossentropy: 0.0246 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0418 1875/1875 [==============================] - ETA: 0s - loss: 0.0663 - sparse_categorical_crossentropy: 0.0246 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0418 1875/1875 [==============================] - 12s 6ms/step - loss: 0.0663 - sparse_categorical_crossentropy: 0.0246 - sparse_categorical_accuracy: 0.9926 - scaled_adversarial_loss: 0.0418 Epoch 5/5 1/1875 [..............................] - ETA: 12:26 - loss: 0.0126 - sparse_categorical_crossentropy: 2.6778e-04 - sparse_categorical_accuracy: 1.0000 - scaled_adversarial_loss: 0.0124 9/1875 [..............................] - ETA: 12s - loss: 0.0611 - sparse_categorical_crossentropy: 0.0128 - sparse_categorical_accuracy: 0.9974 - scaled_adversarial_loss: 0.0483 17/1875 [..............................] - ETA: 12s - loss: 0.0659 - sparse_categorical_crossentropy: 0.0167 - sparse_categorical_accuracy: 0.9967 - scaled_adversarial_loss: 0.0492 25/1875 [..............................] - ETA: 12s - loss: 0.0703 - sparse_categorical_crossentropy: 0.0204 - sparse_categorical_accuracy: 0.9953 - scaled_adversarial_loss: 0.0500 34/1875 [..............................] - ETA: 11s - loss: 0.0723 - sparse_categorical_crossentropy: 0.0216 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0507 43/1875 [..............................] - ETA: 11s - loss: 0.0727 - sparse_categorical_crossentropy: 0.0219 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0507 52/1875 [..............................] - ETA: 11s - loss: 0.0729 - sparse_categorical_crossentropy: 0.0223 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0506 61/1875 [..............................] - ETA: 11s - loss: 0.0733 - sparse_categorical_crossentropy: 0.0227 - sparse_categorical_accuracy: 0.9941 - scaled_adversarial_loss: 0.0506 69/1875 [>.............................] - ETA: 11s - loss: 0.0735 - sparse_categorical_crossentropy: 0.0229 - sparse_categorical_accuracy: 0.9940 - scaled_adversarial_loss: 0.0506 77/1875 [>.............................] - ETA: 11s - loss: 0.0733 - sparse_categorical_crossentropy: 0.0228 - sparse_categorical_accuracy: 0.9940 - scaled_adversarial_loss: 0.0504 85/1875 [>.............................] - ETA: 11s - loss: 0.0729 - sparse_categorical_crossentropy: 0.0227 - sparse_categorical_accuracy: 0.9940 - scaled_adversarial_loss: 0.0502 94/1875 [>.............................] - ETA: 11s - loss: 0.0723 - sparse_categorical_crossentropy: 0.0226 - sparse_categorical_accuracy: 0.9940 - scaled_adversarial_loss: 0.0497 102/1875 [>.............................] - ETA: 11s - loss: 0.0716 - sparse_categorical_crossentropy: 0.0225 - sparse_categorical_accuracy: 0.9940 - scaled_adversarial_loss: 0.0492 110/1875 [>.............................] - ETA: 11s - loss: 0.0708 - sparse_categorical_crossentropy: 0.0223 - sparse_categorical_accuracy: 0.9940 - scaled_adversarial_loss: 0.0486 118/1875 [>.............................] - ETA: 11s - loss: 0.0700 - sparse_categorical_crossentropy: 0.0221 - sparse_categorical_accuracy: 0.9941 - scaled_adversarial_loss: 0.0479 126/1875 [=>............................] - ETA: 11s - loss: 0.0692 - sparse_categorical_crossentropy: 0.0219 - sparse_categorical_accuracy: 0.9941 - scaled_adversarial_loss: 0.0473 134/1875 [=>............................] - ETA: 11s - loss: 0.0684 - sparse_categorical_crossentropy: 0.0217 - sparse_categorical_accuracy: 0.9941 - scaled_adversarial_loss: 0.0466 143/1875 [=>............................] - ETA: 11s - loss: 0.0675 - sparse_categorical_crossentropy: 0.0215 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0459 151/1875 [=>............................] - ETA: 11s - loss: 0.0667 - sparse_categorical_crossentropy: 0.0213 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0453 159/1875 [=>............................] - ETA: 11s - loss: 0.0659 - sparse_categorical_crossentropy: 0.0211 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0448 168/1875 [=>............................] - ETA: 10s - loss: 0.0652 - sparse_categorical_crossentropy: 0.0209 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0443 176/1875 [=>............................] - ETA: 10s - loss: 0.0646 - sparse_categorical_crossentropy: 0.0208 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0439 184/1875 [=>............................] - ETA: 10s - loss: 0.0642 - sparse_categorical_crossentropy: 0.0206 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0436 193/1875 [==>...........................] - ETA: 10s - loss: 0.0638 - sparse_categorical_crossentropy: 0.0205 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0433 202/1875 [==>...........................] - ETA: 10s - loss: 0.0635 - sparse_categorical_crossentropy: 0.0203 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0431 211/1875 [==>...........................] - ETA: 10s - loss: 0.0632 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0429 219/1875 [==>...........................] - ETA: 10s - loss: 0.0629 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0428 227/1875 [==>...........................] - ETA: 10s - loss: 0.0627 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0427 235/1875 [==>...........................] - ETA: 10s - loss: 0.0625 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0426 243/1875 [==>...........................] - ETA: 10s - loss: 0.0624 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0425 251/1875 [===>..........................] - ETA: 10s - loss: 0.0623 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0424 259/1875 [===>..........................] - ETA: 10s - loss: 0.0622 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0424 267/1875 [===>..........................] - ETA: 10s - loss: 0.0621 - sparse_categorical_crossentropy: 0.0197 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0424 275/1875 [===>..........................] - ETA: 10s - loss: 0.0620 - sparse_categorical_crossentropy: 0.0197 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0423 282/1875 [===>..........................] - ETA: 10s - loss: 0.0620 - sparse_categorical_crossentropy: 0.0197 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0423 290/1875 [===>..........................] - ETA: 10s - loss: 0.0619 - sparse_categorical_crossentropy: 0.0197 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0422 298/1875 [===>..........................] - ETA: 10s - loss: 0.0619 - sparse_categorical_crossentropy: 0.0197 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0422 306/1875 [===>..........................] - ETA: 10s - loss: 0.0620 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0422 315/1875 [====>.........................] - ETA: 10s - loss: 0.0620 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0421 324/1875 [====>.........................] - ETA: 10s - loss: 0.0619 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0421 332/1875 [====>.........................] - ETA: 9s - loss: 0.0619 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0420 340/1875 [====>.........................] - ETA: 9s - loss: 0.0619 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0419 349/1875 [====>.........................] - ETA: 9s - loss: 0.0618 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0419 358/1875 [====>.........................] - ETA: 9s - loss: 0.0617 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0418 366/1875 [====>.........................] - ETA: 9s - loss: 0.0617 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0417 374/1875 [====>.........................] - ETA: 9s - loss: 0.0616 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0416 382/1875 [=====>........................] - ETA: 9s - loss: 0.0615 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0415 390/1875 [=====>........................] - ETA: 9s - loss: 0.0615 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9947 - scaled_adversarial_loss: 0.0414 398/1875 [=====>........................] - ETA: 9s - loss: 0.0614 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0413 407/1875 [=====>........................] - ETA: 9s - loss: 0.0613 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0412 415/1875 [=====>........................] - ETA: 9s - loss: 0.0612 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0411 423/1875 [=====>........................] - ETA: 9s - loss: 0.0612 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0410 431/1875 [=====>........................] - ETA: 9s - loss: 0.0611 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0409 439/1875 [======>.......................] - ETA: 9s - loss: 0.0610 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0408 446/1875 [======>.......................] - ETA: 9s - loss: 0.0609 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0407 454/1875 [======>.......................] - ETA: 9s - loss: 0.0608 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0406 462/1875 [======>.......................] - ETA: 9s - loss: 0.0607 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0405 471/1875 [======>.......................] - ETA: 9s - loss: 0.0606 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0404 480/1875 [======>.......................] - ETA: 9s - loss: 0.0606 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0404 488/1875 [======>.......................] - ETA: 8s - loss: 0.0605 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9946 - scaled_adversarial_loss: 0.0403 496/1875 [======>.......................] - ETA: 8s - loss: 0.0605 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0402 504/1875 [=======>......................] - ETA: 8s - loss: 0.0604 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0402 513/1875 [=======>......................] - ETA: 8s - loss: 0.0604 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0402 522/1875 [=======>......................] - ETA: 8s - loss: 0.0604 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0401 531/1875 [=======>......................] - ETA: 8s - loss: 0.0603 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0401 539/1875 [=======>......................] - ETA: 8s - loss: 0.0603 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0401 547/1875 [=======>......................] - ETA: 8s - loss: 0.0603 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0401 555/1875 [=======>......................] - ETA: 8s - loss: 0.0603 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0401 564/1875 [========>.....................] - ETA: 8s - loss: 0.0603 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0400 572/1875 [========>.....................] - ETA: 8s - loss: 0.0603 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0400 581/1875 [========>.....................] - ETA: 8s - loss: 0.0602 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0400 590/1875 [========>.....................] - ETA: 8s - loss: 0.0602 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0400 598/1875 [========>.....................] - ETA: 8s - loss: 0.0602 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0399 606/1875 [========>.....................] - ETA: 8s - loss: 0.0601 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0399 614/1875 [========>.....................] - ETA: 8s - loss: 0.0601 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0399 623/1875 [========>.....................] - ETA: 8s - loss: 0.0601 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0399 631/1875 [=========>....................] - ETA: 8s - loss: 0.0600 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0398 639/1875 [=========>....................] - ETA: 7s - loss: 0.0600 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0398 647/1875 [=========>....................] - ETA: 7s - loss: 0.0600 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0398 655/1875 [=========>....................] - ETA: 7s - loss: 0.0599 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0397 663/1875 [=========>....................] - ETA: 7s - loss: 0.0599 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0397 672/1875 [=========>....................] - ETA: 7s - loss: 0.0598 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0397 681/1875 [=========>....................] - ETA: 7s - loss: 0.0598 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0396 689/1875 [==========>...................] - ETA: 7s - loss: 0.0598 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0396 697/1875 [==========>...................] - ETA: 7s - loss: 0.0597 - sparse_categorical_crossentropy: 0.0202 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0396 705/1875 [==========>...................] - ETA: 7s - loss: 0.0597 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0395 713/1875 [==========>...................] - ETA: 7s - loss: 0.0596 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0395 721/1875 [==========>...................] - ETA: 7s - loss: 0.0596 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0394 729/1875 [==========>...................] - ETA: 7s - loss: 0.0595 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0394 738/1875 [==========>...................] - ETA: 7s - loss: 0.0595 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0394 747/1875 [==========>...................] - ETA: 7s - loss: 0.0594 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0393 756/1875 [===========>..................] - ETA: 7s - loss: 0.0594 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0393 764/1875 [===========>..................] - ETA: 7s - loss: 0.0593 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0392 772/1875 [===========>..................] - ETA: 7s - loss: 0.0592 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0392 781/1875 [===========>..................] - ETA: 7s - loss: 0.0592 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0391 789/1875 [===========>..................] - ETA: 7s - loss: 0.0591 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0391 797/1875 [===========>..................] - ETA: 6s - loss: 0.0591 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0391 806/1875 [===========>..................] - ETA: 6s - loss: 0.0590 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9945 - scaled_adversarial_loss: 0.0390 814/1875 [============>.................] - ETA: 6s - loss: 0.0590 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0390 822/1875 [============>.................] - ETA: 6s - loss: 0.0589 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0389 831/1875 [============>.................] - ETA: 6s - loss: 0.0589 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0389 839/1875 [============>.................] - ETA: 6s - loss: 0.0589 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0389 847/1875 [============>.................] - ETA: 6s - loss: 0.0588 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0388 855/1875 [============>.................] - ETA: 6s - loss: 0.0588 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0388 863/1875 [============>.................] - ETA: 6s - loss: 0.0588 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0388 871/1875 [============>.................] - ETA: 6s - loss: 0.0588 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0387 879/1875 [=============>................] - ETA: 6s - loss: 0.0587 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0387 887/1875 [=============>................] - ETA: 6s - loss: 0.0587 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0387 895/1875 [=============>................] - ETA: 6s - loss: 0.0587 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0386 903/1875 [=============>................] - ETA: 6s - loss: 0.0586 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0386 912/1875 [=============>................] - ETA: 6s - loss: 0.0586 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0386 920/1875 [=============>................] - ETA: 6s - loss: 0.0586 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0385 928/1875 [=============>................] - ETA: 6s - loss: 0.0586 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0385 936/1875 [=============>................] - ETA: 6s - loss: 0.0585 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0385 945/1875 [==============>...............] - ETA: 5s - loss: 0.0585 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0384 954/1875 [==============>...............] - ETA: 5s - loss: 0.0585 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0384 963/1875 [==============>...............] - ETA: 5s - loss: 0.0584 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0383 972/1875 [==============>...............] - ETA: 5s - loss: 0.0584 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0383 980/1875 [==============>...............] - ETA: 5s - loss: 0.0584 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9944 - scaled_adversarial_loss: 0.0383 988/1875 [==============>...............] - ETA: 5s - loss: 0.0583 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0382 997/1875 [==============>...............] - ETA: 5s - loss: 0.0583 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0382 1005/1875 [===============>..............] - ETA: 5s - loss: 0.0583 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0382 1014/1875 [===============>..............] - ETA: 5s - loss: 0.0582 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0381 1023/1875 [===============>..............] - ETA: 5s - loss: 0.0582 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0381 1031/1875 [===============>..............] - ETA: 5s - loss: 0.0582 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0380 1039/1875 [===============>..............] - ETA: 5s - loss: 0.0581 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0380 1047/1875 [===============>..............] - ETA: 5s - loss: 0.0581 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0380 1056/1875 [===============>..............] - ETA: 5s - loss: 0.0581 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0379 1065/1875 [================>.............] - ETA: 5s - loss: 0.0580 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0379 1074/1875 [================>.............] - ETA: 5s - loss: 0.0580 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0379 1082/1875 [================>.............] - ETA: 5s - loss: 0.0579 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0378 1090/1875 [================>.............] - ETA: 5s - loss: 0.0579 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0378 1098/1875 [================>.............] - ETA: 4s - loss: 0.0579 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0378 1107/1875 [================>.............] - ETA: 4s - loss: 0.0578 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0377 1115/1875 [================>.............] - ETA: 4s - loss: 0.0578 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0377 1124/1875 [================>.............] - ETA: 4s - loss: 0.0577 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0376 1132/1875 [=================>............] - ETA: 4s - loss: 0.0577 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0376 1140/1875 [=================>............] - ETA: 4s - loss: 0.0577 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0376 1148/1875 [=================>............] - ETA: 4s - loss: 0.0576 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0375 1156/1875 [=================>............] - ETA: 4s - loss: 0.0576 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0375 1165/1875 [=================>............] - ETA: 4s - loss: 0.0576 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0375 1173/1875 [=================>............] - ETA: 4s - loss: 0.0576 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0375 1182/1875 [=================>............] - ETA: 4s - loss: 0.0575 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0374 1190/1875 [==================>...........] - ETA: 4s - loss: 0.0575 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0374 1198/1875 [==================>...........] - ETA: 4s - loss: 0.0575 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9943 - scaled_adversarial_loss: 0.0374 1207/1875 [==================>...........] - ETA: 4s - loss: 0.0575 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0374 1215/1875 [==================>...........] - ETA: 4s - loss: 0.0574 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0373 1223/1875 [==================>...........] - ETA: 4s - loss: 0.0574 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0373 1231/1875 [==================>...........] - ETA: 4s - loss: 0.0574 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0373 1239/1875 [==================>...........] - ETA: 4s - loss: 0.0573 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0373 1247/1875 [==================>...........] - ETA: 4s - loss: 0.0573 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0372 1255/1875 [===================>..........] - ETA: 3s - loss: 0.0573 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0372 1264/1875 [===================>..........] - ETA: 3s - loss: 0.0573 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0372 1272/1875 [===================>..........] - ETA: 3s - loss: 0.0572 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0372 1280/1875 [===================>..........] - ETA: 3s - loss: 0.0572 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0371 1288/1875 [===================>..........] - ETA: 3s - loss: 0.0572 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0371 1296/1875 [===================>..........] - ETA: 3s - loss: 0.0571 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0371 1304/1875 [===================>..........] - ETA: 3s - loss: 0.0571 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0371 1312/1875 [===================>..........] - ETA: 3s - loss: 0.0571 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0370 1320/1875 [====================>.........] - ETA: 3s - loss: 0.0570 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0370 1329/1875 [====================>.........] - ETA: 3s - loss: 0.0570 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0370 1337/1875 [====================>.........] - ETA: 3s - loss: 0.0570 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0369 1345/1875 [====================>.........] - ETA: 3s - loss: 0.0569 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0369 1353/1875 [====================>.........] - ETA: 3s - loss: 0.0569 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0369 1361/1875 [====================>.........] - ETA: 3s - loss: 0.0569 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0368 1369/1875 [====================>.........] - ETA: 3s - loss: 0.0568 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0368 1378/1875 [=====================>........] - ETA: 3s - loss: 0.0568 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0368 1387/1875 [=====================>........] - ETA: 3s - loss: 0.0567 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0368 1396/1875 [=====================>........] - ETA: 3s - loss: 0.0567 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0367 1404/1875 [=====================>........] - ETA: 3s - loss: 0.0567 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0367 1412/1875 [=====================>........] - ETA: 2s - loss: 0.0566 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0367 1420/1875 [=====================>........] - ETA: 2s - loss: 0.0566 - sparse_categorical_crossentropy: 0.0200 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0366 1428/1875 [=====================>........] - ETA: 2s - loss: 0.0565 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0366 1436/1875 [=====================>........] - ETA: 2s - loss: 0.0565 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0366 1444/1875 [======================>.......] - ETA: 2s - loss: 0.0565 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0365 1452/1875 [======================>.......] - ETA: 2s - loss: 0.0564 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0365 1459/1875 [======================>.......] - ETA: 2s - loss: 0.0564 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0365 1466/1875 [======================>.......] - ETA: 2s - loss: 0.0564 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0365 1474/1875 [======================>.......] - ETA: 2s - loss: 0.0563 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0364 1482/1875 [======================>.......] - ETA: 2s - loss: 0.0563 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0364 1490/1875 [======================>.......] - ETA: 2s - loss: 0.0563 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0364 1498/1875 [======================>.......] - ETA: 2s - loss: 0.0563 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0364 1506/1875 [=======================>......] - ETA: 2s - loss: 0.0562 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0363 1514/1875 [=======================>......] - ETA: 2s - loss: 0.0562 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0363 1522/1875 [=======================>......] - ETA: 2s - loss: 0.0562 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0363 1530/1875 [=======================>......] - ETA: 2s - loss: 0.0562 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0363 1539/1875 [=======================>......] - ETA: 2s - loss: 0.0561 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0363 1547/1875 [=======================>......] - ETA: 2s - loss: 0.0561 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0362 1555/1875 [=======================>......] - ETA: 2s - loss: 0.0561 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0362 1563/1875 [========================>.....] - ETA: 2s - loss: 0.0561 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0362 1573/1875 [========================>.....] - ETA: 1s - loss: 0.0561 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0362 1583/1875 [========================>.....] - ETA: 1s - loss: 0.0560 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0362 1594/1875 [========================>.....] - ETA: 1s - loss: 0.0560 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0362 1605/1875 [========================>.....] - ETA: 1s - loss: 0.0560 - sparse_categorical_crossentropy: 0.0199 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0361 1616/1875 [========================>.....] - ETA: 1s - loss: 0.0560 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0361 1625/1875 [=========================>....] - ETA: 1s - loss: 0.0559 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0361 1634/1875 [=========================>....] - ETA: 1s - loss: 0.0559 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0361 1644/1875 [=========================>....] - ETA: 1s - loss: 0.0559 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0360 1654/1875 [=========================>....] - ETA: 1s - loss: 0.0559 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0360 1665/1875 [=========================>....] - ETA: 1s - loss: 0.0558 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0360 1675/1875 [=========================>....] - ETA: 1s - loss: 0.0558 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0360 1686/1875 [=========================>....] - ETA: 1s - loss: 0.0558 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0359 1696/1875 [==========================>...] - ETA: 1s - loss: 0.0557 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0359 1706/1875 [==========================>...] - ETA: 1s - loss: 0.0557 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0359 1716/1875 [==========================>...] - ETA: 1s - loss: 0.0557 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0359 1726/1875 [==========================>...] - ETA: 0s - loss: 0.0556 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0358 1737/1875 [==========================>...] - ETA: 0s - loss: 0.0556 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0358 1748/1875 [==========================>...] - ETA: 0s - loss: 0.0556 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0358 1758/1875 [===========================>..] - ETA: 0s - loss: 0.0556 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0358 1768/1875 [===========================>..] - ETA: 0s - loss: 0.0555 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0358 1778/1875 [===========================>..] - ETA: 0s - loss: 0.0555 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0357 1788/1875 [===========================>..] - ETA: 0s - loss: 0.0555 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0357 1798/1875 [===========================>..] - ETA: 0s - loss: 0.0554 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0357 1808/1875 [===========================>..] - ETA: 0s - loss: 0.0554 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0357 1819/1875 [============================>.] - ETA: 0s - loss: 0.0554 - sparse_categorical_crossentropy: 0.0198 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0356 1829/1875 [============================>.] - ETA: 0s - loss: 0.0554 - sparse_categorical_crossentropy: 0.0197 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0356 1839/1875 [============================>.] - ETA: 0s - loss: 0.0554 - sparse_categorical_crossentropy: 0.0197 - sparse_categorical_accuracy: 0.9942 - scaled_adversarial_loss: 0.0356 1849/1875 [============================>.] - ETA: 0s - loss: 0.0553 - sparse_categorical_crossentropy: 0.0197 - sparse_categorical_accuracy: 0.9941 - scaled_adversarial_loss: 0.0356 1860/1875 [============================>.] - ETA: 0s - loss: 0.0553 - sparse_categorical_crossentropy: 0.0197 - sparse_categorical_accuracy: 0.9941 - scaled_adversarial_loss: 0.0356 1871/1875 [============================>.] - ETA: 0s - loss: 0.0553 - sparse_categorical_crossentropy: 0.0197 - sparse_categorical_accuracy: 0.9941 - scaled_adversarial_loss: 0.0356 1875/1875 [==============================] - 12s 6ms/step - loss: 0.0553 - sparse_categorical_crossentropy: 0.0197 - sparse_categorical_accuracy: 0.9941 - scaled_adversarial_loss: 0.0355 <tensorflow.python.keras.callbacks.History at 0x7fa32c9e0240>
results = adv_model.evaluate(test_set_for_adv_model)
named_results = dict(zip(adv_model.metrics_names, results))
print('\naccuracy:', named_results['sparse_categorical_accuracy'])
WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 1/313 [..............................] - ETA: 1:27 - loss: 0.0236 - sparse_categorical_crossentropy: 0.0028 - sparse_categorical_accuracy: 1.0000 - scaled_adversarial_loss: 0.0208 14/313 [>.............................] - ETA: 1s - loss: 0.0627 - sparse_categorical_crossentropy: 0.0324 - sparse_categorical_accuracy: 0.9955 - scaled_adversarial_loss: 0.0303 28/313 [=>............................] - ETA: 1s - loss: 0.0603 - sparse_categorical_crossentropy: 0.0295 - sparse_categorical_accuracy: 0.9933 - scaled_adversarial_loss: 0.0308 41/313 [==>...........................] - ETA: 1s - loss: 0.0558 - sparse_categorical_crossentropy: 0.0226 - sparse_categorical_accuracy: 0.9954 - scaled_adversarial_loss: 0.0332 54/313 [====>.........................] - ETA: 1s - loss: 0.0532 - sparse_categorical_crossentropy: 0.0212 - sparse_categorical_accuracy: 0.9948 - scaled_adversarial_loss: 0.0320 67/313 [=====>........................] - ETA: 0s - loss: 0.0512 - sparse_categorical_crossentropy: 0.0192 - sparse_categorical_accuracy: 0.9949 - scaled_adversarial_loss: 0.0320 80/313 [======>.......................] - ETA: 0s - loss: 0.0533 - sparse_categorical_crossentropy: 0.0220 - sparse_categorical_accuracy: 0.9941 - scaled_adversarial_loss: 0.0313 93/313 [=======>......................] - ETA: 0s - loss: 0.0555 - sparse_categorical_crossentropy: 0.0225 - sparse_categorical_accuracy: 0.9936 - scaled_adversarial_loss: 0.0329 106/313 [=========>....................] - ETA: 0s - loss: 0.0549 - sparse_categorical_crossentropy: 0.0219 - sparse_categorical_accuracy: 0.9932 - scaled_adversarial_loss: 0.0330 119/313 [==========>...................] - ETA: 0s - loss: 0.0566 - sparse_categorical_crossentropy: 0.0232 - sparse_categorical_accuracy: 0.9932 - scaled_adversarial_loss: 0.0333 132/313 [===========>..................] - ETA: 0s - loss: 0.0558 - sparse_categorical_crossentropy: 0.0228 - sparse_categorical_accuracy: 0.9934 - scaled_adversarial_loss: 0.0330 144/313 [============>.................] - ETA: 0s - loss: 0.0547 - sparse_categorical_crossentropy: 0.0222 - sparse_categorical_accuracy: 0.9935 - scaled_adversarial_loss: 0.0325 156/313 [=============>................] - ETA: 0s - loss: 0.0544 - sparse_categorical_crossentropy: 0.0228 - sparse_categorical_accuracy: 0.9930 - scaled_adversarial_loss: 0.0316 169/313 [===============>..............] - ETA: 0s - loss: 0.0538 - sparse_categorical_crossentropy: 0.0222 - sparse_categorical_accuracy: 0.9930 - scaled_adversarial_loss: 0.0317 182/313 [================>.............] - ETA: 0s - loss: 0.0533 - sparse_categorical_crossentropy: 0.0214 - sparse_categorical_accuracy: 0.9933 - scaled_adversarial_loss: 0.0319 195/313 [=================>............] - ETA: 0s - loss: 0.0526 - sparse_categorical_crossentropy: 0.0210 - sparse_categorical_accuracy: 0.9936 - scaled_adversarial_loss: 0.0316 208/313 [==================>...........] - ETA: 0s - loss: 0.0523 - sparse_categorical_crossentropy: 0.0208 - sparse_categorical_accuracy: 0.9934 - scaled_adversarial_loss: 0.0315 220/313 [====================>.........] - ETA: 0s - loss: 0.0514 - sparse_categorical_crossentropy: 0.0201 - sparse_categorical_accuracy: 0.9935 - scaled_adversarial_loss: 0.0314 232/313 [=====================>........] - ETA: 0s - loss: 0.0531 - sparse_categorical_crossentropy: 0.0213 - sparse_categorical_accuracy: 0.9931 - scaled_adversarial_loss: 0.0319 245/313 [======================>.......] - ETA: 0s - loss: 0.0535 - sparse_categorical_crossentropy: 0.0219 - sparse_categorical_accuracy: 0.9929 - scaled_adversarial_loss: 0.0315 258/313 [=======================>......] - ETA: 0s - loss: 0.0546 - sparse_categorical_crossentropy: 0.0226 - sparse_categorical_accuracy: 0.9925 - scaled_adversarial_loss: 0.0321 271/313 [========================>.....] - ETA: 0s - loss: 0.0558 - sparse_categorical_crossentropy: 0.0236 - sparse_categorical_accuracy: 0.9923 - scaled_adversarial_loss: 0.0322 284/313 [==========================>...] - ETA: 0s - loss: 0.0566 - sparse_categorical_crossentropy: 0.0243 - sparse_categorical_accuracy: 0.9922 - scaled_adversarial_loss: 0.0324 297/313 [===========================>..] - ETA: 0s - loss: 0.0569 - sparse_categorical_crossentropy: 0.0244 - sparse_categorical_accuracy: 0.9922 - scaled_adversarial_loss: 0.0326 310/313 [============================>.] - ETA: 0s - loss: 0.0567 - sparse_categorical_crossentropy: 0.0242 - sparse_categorical_accuracy: 0.9922 - scaled_adversarial_loss: 0.0326 313/313 [==============================] - 2s 4ms/step - loss: 0.0566 - sparse_categorical_crossentropy: 0.0239 - sparse_categorical_accuracy: 0.9923 - scaled_adversarial_loss: 0.0326 accuracy: 0.9922999739646912
敵対的正則化モデルもまた、テストセットで非常に優れた性能（99% の精度）を達成していることが分かります。
敵対的摂動下におけるロバスト性
ここでは基本モデルと敵対的正則化モデルの、敵対的摂動下におけるロバスト性を比較します。
敵対的摂動の例の生成には AdversarialRegularization.perturb_on_batch 関数を使用します。そして、これは基本モデルを基にして生成します。これを行うためには、AdversarialRegularization を使用して基本モデルをラップします。トレーニング（Model.fit）を呼び出さない限り、モデル内の学習変数に変更はなく、モデルは基本モデルのセクションのものと同じであることに注意してください。
reference_model = nsl.keras.AdversarialRegularization(
base_model,
label_keys=[LABEL_INPUT_NAME],
adv_config=adv_config)
reference_model.compile(
optimizer='adam',
loss='sparse_categorical_crossentropy',
metrics=['acc'])
評価対象のモデルをディクショナリに収集し、各モデルのメトリックオブジェクトを作成します。
基本モデルと同じ（ラベル情報が不要な）入力形式を持つために adv_model.base_model を取る必要があることに注意してください。adv_model.base_model で学習する変数は、adv_model で学習する変数と同じです。
models_to_eval = {
'base': base_model,
'adv-regularized': adv_model.base_model
}
metrics = {
name: tf.keras.metrics.SparseCategoricalAccuracy()
for name in models_to_eval.keys()
}
摂動された例を生成し、それを用いてモデルを評価するループをここに示します。摂動された画像、ラベル、予測は次のセクションで可視化するために保存しておきます。
perturbed_images, labels, predictions = [], [], []
for batch in test_set_for_adv_model:
perturbed_batch = reference_model.perturb_on_batch(batch)
# Clipping makes perturbed examples have the same range as regular ones.
perturbed_batch[IMAGE_INPUT_NAME] = tf.clip_by_value(
perturbed_batch[IMAGE_INPUT_NAME], 0.0, 1.0)
y_true = perturbed_batch.pop(LABEL_INPUT_NAME)
perturbed_images.append(perturbed_batch[IMAGE_INPUT_NAME].numpy())
labels.append(y_true.numpy())
predictions.append({})
for name, model in models_to_eval.items():
y_pred = model(perturbed_batch)
metrics[name](y_true, y_pred)
predictions[-1][name] = tf.argmax(y_pred, axis=-1).numpy()
for name, metric in metrics.items():
print('%s model accuracy: %f' % (name, metric.result().numpy()))
WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:absl:Cannot perturb features dict_keys(['label']) WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the watched tensor must be floating (e.g. tf.float32), got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 WARNING:tensorflow:The dtype of the source tensor must be floating (e.g. tf.float32) when calling GradientTape.gradient, got tf.int64 base model accuracy: 0.406500 adv-regularized model accuracy: 0.955700
入力が敵対的に摂動されると、基本モデルの精度が劇的に（99% から約 50% に）低下することが分かります。一方で、敵対的正則化されたモデルの精度低下はごくわずか（99% から 95% に）です。これは、敵対的学習がモデルのロバスト性向上に有効であることを示しています。
敵対的摂動された画像の例
ここで、敵対的摂動された画像を確認してみます。摂動された画像は人間が認識可能な数字を表示していますが、基本モデルをうまく騙せることが分かります。
batch_index = 0
batch_image = perturbed_images[batch_index]
batch_label = labels[batch_index]
batch_pred = predictions[batch_index]
batch_size = HPARAMS.batch_size
n_col = 4
n_row = (batch_size + n_col - 1) / n_col
print('accuracy in batch %d:' % batch_index)
for name, pred in batch_pred.items():
print('%s model: %d / %d' % (name, np.sum(batch_label == pred), batch_size))
plt.figure(figsize=(15, 15))
for i, (image, y) in enumerate(zip(batch_image, batch_label)):
y_base = batch_pred['base'][i]
y_adv = batch_pred['adv-regularized'][i]
plt.subplot(n_row, n_col, i+1)
plt.title('true: %d, base: %d, adv: %d' % (y, y_base, y_adv))
plt.imshow(tf.keras.preprocessing.image.array_to_img(image), cmap='gray')
plt.axis('off')
plt.show()
accuracy in batch 0: base model: 15 / 32 adv-regularized model: 30 / 32 /home/kbuilder/.local/lib/python3.6/site-packages/ipykernel_launcher.py:19: MatplotlibDeprecationWarning: Passing non-integers as three-element position specification is deprecated since 3.3 and will be removed two minor releases later.
結論
Neural Structured Learning（NSL）フレームワークを使用して画像分類に敵対的学習を用いることを実証しました。ユーザーの方々には、（ハイパーパラメータの）さまざまな敵対的設定を実験し、それがモデルのロバスト性にどのような影響を与えるかを確認してみることを推奨してしています。