 View on TensorFlow.org
|
 Run in Google Colab
|
 View source on GitHub
|
 Download notebook
|
Overview
The tf.distribute.Strategy API provides an abstraction for distributing your training
across multiple processing units. The goal is to allow users to enable distributed training using existing models and training code, with minimal changes.
This tutorial uses the tf.distribute.MirroredStrategy, which
does in-graph replication with synchronous training on many GPUs on one machine.
Essentially, it copies all of the model's variables to each processor.
Then, it uses all-reduce to combine the gradients from all processors and applies the combined value to all copies of the model.
MirroredStrategy is one of several distribution strategy available in TensorFlow core. You can read about more strategies at distribution strategy guide.
Keras API
This example uses the tf.keras API to build the model and training loop. For custom training loops, see the tf.distribute.Strategy with training loops tutorial.
Import dependencies
from __future__ import absolute_import, division, print_function, unicode_literals
# Import TensorFlow and TensorFlow Datasets
try:
!pip install -q tf-nightly
except Exception:
pass
import tensorflow_datasets as tfds
import tensorflow as tf
tfds.disable_progress_bar()
import os
print(tf.__version__)
2.1.0-dev20191118
Download the dataset
Download the MNIST dataset and load it from TensorFlow Datasets. This returns a dataset in tf.data format.
Setting with_info to True includes the metadata for the entire dataset, which is being saved here to info.
Among other things, this metadata object includes the number of train and test examples.
datasets, info = tfds.load(name='mnist', with_info=True, as_supervised=True)
mnist_train, mnist_test = datasets['train'], datasets['test']
Downloading and preparing dataset mnist (11.06 MiB) to /home/kbuilder/tensorflow_datasets/mnist/1.0.0... WARNING:absl:Dataset mnist is hosted on GCS. It will automatically be downloaded to your local data directory. If you'd instead prefer to read directly from our public GCS bucket (recommended if you're running on GCP), you can instead set data_dir=gs://tfds-data/datasets. Dataset mnist downloaded and prepared to /home/kbuilder/tensorflow_datasets/mnist/1.0.0. Subsequent calls will reuse this data.
Define distribution strategy
Create a MirroredStrategy object. This will handle distribution, and provides a context manager (tf.distribute.MirroredStrategy.scope) to build your model inside.
strategy = tf.distribute.MirroredStrategy()
WARNING:tensorflow:There are non-GPU devices in `tf.distribute.Strategy`, not using nccl allreduce.
WARNING:tensorflow:There are non-GPU devices in `tf.distribute.Strategy`, not using nccl allreduce.
INFO:tensorflow:Using MirroredStrategy with devices ('/job:localhost/replica:0/task:0/device:CPU:0',)
INFO:tensorflow:Using MirroredStrategy with devices ('/job:localhost/replica:0/task:0/device:CPU:0',)
print('Number of devices: {}'.format(strategy.num_replicas_in_sync))
Number of devices: 1
Setup input pipeline
When training a model with multiple GPUs, you can use the extra computing power effectively by increasing the batch size. In general, use the largest batch size that fits the GPU memory, and tune the learning rate accordingly.
# You can also do info.splits.total_num_examples to get the total
# number of examples in the dataset.
num_train_examples = info.splits['train'].num_examples
num_test_examples = info.splits['test'].num_examples
BUFFER_SIZE = 10000
BATCH_SIZE_PER_REPLICA = 64
BATCH_SIZE = BATCH_SIZE_PER_REPLICA * strategy.num_replicas_in_sync
Pixel values, which are 0-255, have to be normalized to the 0-1 range. Define this scale in a function.
def scale(image, label):
image = tf.cast(image, tf.float32)
image /= 255
return image, label
Apply this function to the training and test data, shuffle the training data, and batch it for training. Notice we are also keeping an in-memory cache of the training data to improve performance.
train_dataset = mnist_train.map(scale).cache().shuffle(BUFFER_SIZE).batch(BATCH_SIZE)
eval_dataset = mnist_test.map(scale).batch(BATCH_SIZE)
Create the model
Create and compile the Keras model in the context of strategy.scope.
with strategy.scope():
model = tf.keras.Sequential([
tf.keras.layers.Conv2D(32, 3, activation='relu', input_shape=(28, 28, 1)),
tf.keras.layers.MaxPooling2D(),
tf.keras.layers.Flatten(),
tf.keras.layers.Dense(64, activation='relu'),
tf.keras.layers.Dense(10, activation='softmax')
])
model.compile(loss='sparse_categorical_crossentropy',
optimizer=tf.keras.optimizers.Adam(),
metrics=['accuracy'])
Define the callbacks
The callbacks used here are:
- TensorBoard: This callback writes a log for TensorBoard which allows you to visualize the graphs.
- Model Checkpoint: This callback saves the model after every epoch.
- Learning Rate Scheduler: Using this callback, you can schedule the learning rate to change after every epoch/batch.
For illustrative purposes, add a print callback to display the learning rate in the notebook.
# Define the checkpoint directory to store the checkpoints
checkpoint_dir = './training_checkpoints'
# Name of the checkpoint files
checkpoint_prefix = os.path.join(checkpoint_dir, "ckpt_{epoch}")
# Function for decaying the learning rate.
# You can define any decay function you need.
def decay(epoch):
if epoch < 3:
return 1e-3
elif epoch >= 3 and epoch < 7:
return 1e-4
else:
return 1e-5
# Callback for printing the LR at the end of each epoch.
class PrintLR(tf.keras.callbacks.Callback):
def on_epoch_end(self, epoch, logs=None):
print('\nLearning rate for epoch {} is {}'.format(epoch + 1,
model.optimizer.lr.numpy()))
callbacks = [
tf.keras.callbacks.TensorBoard(log_dir='./logs'),
tf.keras.callbacks.ModelCheckpoint(filepath=checkpoint_prefix,
save_weights_only=True),
tf.keras.callbacks.LearningRateScheduler(decay),
PrintLR()
]
Train and evaluate
Now, train the model in the usual way, calling fit on the model and passing in the dataset created at the beginning of the tutorial. This step is the same whether you are distributing the training or not.
model.fit(train_dataset, epochs=12, callbacks=callbacks)
Epoch 1/12
938/Unknown - 15s 15ms/step - loss: 0.1910 - accuracy: 0.9443
Learning rate for epoch 1 is 0.0010000000474974513
938/938 [==============================] - 15s 15ms/step - loss: 0.1910 - accuracy: 0.9443
Epoch 2/12
935/938 [============================>.] - ETA: 0s - loss: 0.0639 - accuracy: 0.9807
Learning rate for epoch 2 is 0.0010000000474974513
938/938 [==============================] - 9s 10ms/step - loss: 0.0639 - accuracy: 0.9807
Epoch 3/12
935/938 [============================>.] - ETA: 0s - loss: 0.0438 - accuracy: 0.9874
Learning rate for epoch 3 is 0.0010000000474974513
938/938 [==============================] - 9s 9ms/step - loss: 0.0439 - accuracy: 0.9874
Epoch 4/12
937/938 [============================>.] - ETA: 0s - loss: 0.0236 - accuracy: 0.9938
Learning rate for epoch 4 is 9.999999747378752e-05
938/938 [==============================] - 9s 9ms/step - loss: 0.0237 - accuracy: 0.9937
Epoch 5/12
937/938 [============================>.] - ETA: 0s - loss: 0.0203 - accuracy: 0.9951
Learning rate for epoch 5 is 9.999999747378752e-05
938/938 [==============================] - 9s 9ms/step - loss: 0.0205 - accuracy: 0.9951
Epoch 6/12
936/938 [============================>.] - ETA: 0s - loss: 0.0185 - accuracy: 0.9956
Learning rate for epoch 6 is 9.999999747378752e-05
938/938 [==============================] - 9s 9ms/step - loss: 0.0185 - accuracy: 0.9956
Epoch 7/12
937/938 [============================>.] - ETA: 0s - loss: 0.0172 - accuracy: 0.9959
Learning rate for epoch 7 is 9.999999747378752e-05
938/938 [==============================] - 9s 9ms/step - loss: 0.0172 - accuracy: 0.9959
Epoch 8/12
937/938 [============================>.] - ETA: 0s - loss: 0.0145 - accuracy: 0.9969
Learning rate for epoch 8 is 9.999999747378752e-06
938/938 [==============================] - 9s 9ms/step - loss: 0.0145 - accuracy: 0.9969
Epoch 9/12
937/938 [============================>.] - ETA: 0s - loss: 0.0143 - accuracy: 0.9969
Learning rate for epoch 9 is 9.999999747378752e-06
938/938 [==============================] - 9s 9ms/step - loss: 0.0143 - accuracy: 0.9969
Epoch 10/12
936/938 [============================>.] - ETA: 0s - loss: 0.0142 - accuracy: 0.9969
Learning rate for epoch 10 is 9.999999747378752e-06
938/938 [==============================] - 9s 9ms/step - loss: 0.0142 - accuracy: 0.9969
Epoch 11/12
934/938 [============================>.] - ETA: 0s - loss: 0.0139 - accuracy: 0.9970
Learning rate for epoch 11 is 9.999999747378752e-06
938/938 [==============================] - 9s 9ms/step - loss: 0.0140 - accuracy: 0.9970
Epoch 12/12
937/938 [============================>.] - ETA: 0s - loss: 0.0138 - accuracy: 0.9971
Learning rate for epoch 12 is 9.999999747378752e-06
938/938 [==============================] - 9s 9ms/step - loss: 0.0138 - accuracy: 0.9971
<tensorflow.python.keras.callbacks.History at 0x7fc2fc11c828>
As you can see below, the checkpoints are getting saved.
# check the checkpoint directory
!ls {checkpoint_dir}
checkpoint ckpt_4.data-00000-of-00001 ckpt_1.data-00000-of-00001 ckpt_4.index ckpt_1.index ckpt_5.data-00000-of-00001 ckpt_10.data-00000-of-00001 ckpt_5.index ckpt_10.index ckpt_6.data-00000-of-00001 ckpt_11.data-00000-of-00001 ckpt_6.index ckpt_11.index ckpt_7.data-00000-of-00001 ckpt_12.data-00000-of-00001 ckpt_7.index ckpt_12.index ckpt_8.data-00000-of-00001 ckpt_2.data-00000-of-00001 ckpt_8.index ckpt_2.index ckpt_9.data-00000-of-00001 ckpt_3.data-00000-of-00001 ckpt_9.index ckpt_3.index
To see how the model perform, load the latest checkpoint and call evaluate on the test data.
Call evaluate as before using appropriate datasets.
model.load_weights(tf.train.latest_checkpoint(checkpoint_dir))
eval_loss, eval_acc = model.evaluate(eval_dataset)
print('Eval loss: {}, Eval Accuracy: {}'.format(eval_loss, eval_acc))
157/Unknown - 3s 16ms/step - loss: 0.0364 - accuracy: 0.9880Eval loss: 0.03637826871045325, Eval Accuracy: 0.9879999756813049
To see the output, you can download and view the TensorBoard logs at the terminal.
$ tensorboard --logdir=path/to/log-directory
!ls -sh ./logs
total 4.0K 4.0K train
Export to SavedModel
Export the graph and the variables to the platform-agnostic SavedModel format. After your model is saved, you can load it with or without the scope.
path = 'saved_model/'
model.save(path, save_format='tf')
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.6/site-packages/tensorflow_core/python/ops/resource_variable_ops.py:1788: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version. Instructions for updating: If using Keras pass *_constraint arguments to layers. WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.6/site-packages/tensorflow_core/python/ops/resource_variable_ops.py:1788: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version. Instructions for updating: If using Keras pass *_constraint arguments to layers. INFO:tensorflow:Assets written to: saved_model/assets INFO:tensorflow:Assets written to: saved_model/assets
Load the model without strategy.scope.
unreplicated_model = tf.keras.models.load_model(path)
unreplicated_model.compile(
loss='sparse_categorical_crossentropy',
optimizer=tf.keras.optimizers.Adam(),
metrics=['accuracy'])
eval_loss, eval_acc = unreplicated_model.evaluate(eval_dataset)
print('Eval loss: {}, Eval Accuracy: {}'.format(eval_loss, eval_acc))
157/Unknown - 1s 9ms/step - loss: 0.0364 - accuracy: 0.9880Eval loss: 0.03637869624498116, Eval Accuracy: 0.9879999756813049
Load the model with strategy.scope.
with strategy.scope():
replicated_model = tf.keras.models.load_model(path)
replicated_model.compile(loss='sparse_categorical_crossentropy',
optimizer=tf.keras.optimizers.Adam(),
metrics=['accuracy'])
eval_loss, eval_acc = replicated_model.evaluate(eval_dataset)
print ('Eval loss: {}, Eval Accuracy: {}'.format(eval_loss, eval_acc))
157/Unknown - 1s 9ms/step - loss: 0.0364 - accuracy: 0.9880Eval loss: 0.03637869624498116, Eval Accuracy: 0.9879999756813049
Examples and Tutorials
Here are some examples for using distribution strategy with keras fit/compile:
1. Transformer example trained using tf.distribute.MirroredStrategy
2. NCF example trained using tf.distribute.MirroredStrategy.
More examples listed in the Distribution strategy guide
Next steps
- Read the distribution strategy guide.
- Read the Distributed Training with Custom Training Loops tutorial.