در سمپوزیوم زنان در ML در 7 دسامبر شرکت کنید هم اکنون ثبت نام کنید

طبقه بندی متن با RNN

با مجموعه‌ها، منظم بمانید ذخیره و دسته‌بندی محتوا براساس اولویت‌های شما.

مشاهده در TensorFlow.org در Google Colab اجرا شود مشاهده منبع در GitHub دانلود دفترچه یادداشت

این آموزش طبقه بندی متن قطارها شبکه های عصبی راجعه در IMDB فیلم بزرگ بررسی مجموعه داده برای تجزیه و تحلیل احساسات.

برپایی

import numpy as np

import tensorflow_datasets as tfds
import tensorflow as tf

tfds.disable_progress_bar()

واردات matplotlib و ایجاد یک تابع کمکی به نمودار رسم:

import matplotlib.pyplot as plt


def plot_graphs(history, metric):
  plt.plot(history.history[metric])
  plt.plot(history.history['val_'+metric], '')
  plt.xlabel("Epochs")
  plt.ylabel(metric)
  plt.legend([metric, 'val_'+metric])

راه اندازی خط لوله ورودی

IMDB بزرگ بررسی فیلم مجموعه داده است یک فایل باینری طبقه بندی مجموعه داده همه به بررسی دارند مثبت یا احساسات منفی.

دانلود مجموعه داده با استفاده از TFDS . را ببینید بارگذاری آموزش متن برای جزئیات بیشتر در مورد چگونگی بارگیری این نوع از اطلاعات به صورت دستی.

dataset, info = tfds.load('imdb_reviews', with_info=True,
                          as_supervised=True)
train_dataset, test_dataset = dataset['train'], dataset['test']

train_dataset.element_spec
(TensorSpec(shape=(), dtype=tf.string, name=None),
 TensorSpec(shape=(), dtype=tf.int64, name=None))

در ابتدا مجموعه داده ای از (متن، جفت برچسب) را برمی گرداند:

for example, label in train_dataset.take(1):
  print('text: ', example.numpy())
  print('label: ', label.numpy())
text:  b"This was an absolutely terrible movie. Don't be lured in by Christopher Walken or Michael Ironside. Both are great actors, but this must simply be their worst role in history. Even their great acting could not redeem this movie's ridiculous storyline. This movie is an early nineties US propaganda piece. The most pathetic scenes were those when the Columbian rebels were making their cases for revolutions. Maria Conchita Alonso appeared phony, and her pseudo-love affair with Walken was nothing but a pathetic emotional plug in a movie that was devoid of any real meaning. I am disappointed that there are movies like this, ruining actor's like Christopher Walken's good name. I could barely sit through it."
label:  0

مختلط بعدی داده ها برای آموزش و ایجاد دسته از این (text, label) جفت:

BUFFER_SIZE = 10000
BATCH_SIZE = 64
train_dataset = train_dataset.shuffle(BUFFER_SIZE).batch(BATCH_SIZE).prefetch(tf.data.AUTOTUNE)
test_dataset = test_dataset.batch(BATCH_SIZE).prefetch(tf.data.AUTOTUNE)
for example, label in train_dataset.take(1):
  print('texts: ', example.numpy()[:3])
  print()
  print('labels: ', label.numpy()[:3])
texts:  [b'This is arguably the worst film I have ever seen, and I have quite an appetite for awful (and good) movies. It could (just) have managed a kind of adolescent humour if it had been consistently tongue-in-cheek --\xc3\xa0 la ROCKY HORROR PICTURE SHOW, which was really very funny. Other movies, like PLAN NINE FROM OUTER SPACE, manage to be funny while (apparently) trying to be serious. As to the acting, it looks like they rounded up brain-dead teenagers and asked them to ad-lib the whole production. Compared to them, Tom Cruise looks like Alec Guinness. There was one decent interpretation -- that of the older ghoul-busting broad on the motorcycle.'
 b"I saw this film in the worst possible circumstance. I'd already missed 15 minutes when I woke up to it on an international flight between Sydney and Seoul. I didn't know what I was watching, I thought maybe it was a movie of the week, but quickly became riveted by the performance of the lead actress playing a young woman who's child had been kidnapped. The premise started taking twist and turns I didn't see coming and by the end credits I was scrambling through the the in-flight guide to figure out what I had just watched. Turns out I was belatedly discovering Do-yeon Jeon who'd won Best Actress at Cannes for the role. I don't know if Secret Sunshine is typical of Korean cinema but I'm off to the DVD store to discover more."
 b"Hello. I am Paul Raddick, a.k.a. Panic Attack of WTAF, Channel 29 in Philadelphia. Let me tell you about this god awful movie that powered on Adam Sandler's film career but was digitized after a short time.<br /><br />Going Overboard is about an aspiring comedian played by Sandler who gets a job on a cruise ship and fails...or so I thought. Sandler encounters babes that like History of the World Part 1 and Rebound. The babes were supposed to be engaged, but, actually, they get executed by Sawtooth, the meanest cannibal the world has ever known. Adam Sandler fared bad in Going Overboard, but fared better in Big Daddy, Billy Madison, and Jen Leone's favorite, 50 First Dates. Man, Drew Barrymore was one hot chick. Spanglish is red hot, Going Overboard ain't Dooley squat! End of file."]

labels:  [0 1 0]

رمزگذار متن را ایجاد کنید

متن خام بارگذاری شده توسط tfds نیاز به پردازش می شود قبل از آن را می توان در یک مدل استفاده می شود. ساده ترین راه برای متن روند برای آموزش با استفاده از TextVectorization لایه. این لایه قابلیت های زیادی دارد، اما این آموزش به رفتار پیش فرض پایبند است.

درست لایه، و با تصویب متن مجموعه داده به لایه .adapt روش:

VOCAB_SIZE = 1000
encoder = tf.keras.layers.TextVectorization(
    max_tokens=VOCAB_SIZE)
encoder.adapt(train_dataset.map(lambda text, label: text))

.adapt متد واژگان لایه. در اینجا 20 توکن اول آمده است. پس از padding و نشانه های ناشناخته، آنها بر اساس فرکانس مرتب می شوند:

vocab = np.array(encoder.get_vocabulary())
vocab[:20]
array(['', '[UNK]', 'the', 'and', 'a', 'of', 'to', 'is', 'in', 'it', 'i',
       'this', 'that', 'br', 'was', 'as', 'for', 'with', 'movie', 'but'],
      dtype='<U14')

هنگامی که واژگان تنظیم شد، لایه می تواند متن را در شاخص ها رمزگذاری کند. تانسورها از شاخص 0 پر به طولانی ترین توالی در دسته (مگر اینکه شما در تنظیم یک ثابت output_sequence_length ):

encoded_example = encoder(example)[:3].numpy()
encoded_example
array([[ 11,   7,   1, ...,   0,   0,   0],
       [ 10, 208,  11, ...,   0,   0,   0],
       [  1,  10, 237, ...,   0,   0,   0]])

با تنظیمات پیش فرض، فرآیند به طور کامل برگشت پذیر نیست. سه دلیل اصلی برای آن وجود دارد:

  1. مقدار پیش فرض برای preprocessing.TextVectorization را standardize استدلال این است "lower_and_strip_punctuation" .
  2. اندازه محدود واژگان و فقدان بک گراند مبتنی بر کاراکتر منجر به برخی نشانه های ناشناخته می شود.
for n in range(3):
  print("Original: ", example[n].numpy())
  print("Round-trip: ", " ".join(vocab[encoded_example[n]]))
  print()
Original:  b'This is arguably the worst film I have ever seen, and I have quite an appetite for awful (and good) movies. It could (just) have managed a kind of adolescent humour if it had been consistently tongue-in-cheek --\xc3\xa0 la ROCKY HORROR PICTURE SHOW, which was really very funny. Other movies, like PLAN NINE FROM OUTER SPACE, manage to be funny while (apparently) trying to be serious. As to the acting, it looks like they rounded up brain-dead teenagers and asked them to ad-lib the whole production. Compared to them, Tom Cruise looks like Alec Guinness. There was one decent interpretation -- that of the older ghoul-busting broad on the motorcycle.'
Round-trip:  this is [UNK] the worst film i have ever seen and i have quite an [UNK] for awful and good movies it could just have [UNK] a kind of [UNK] [UNK] if it had been [UNK] [UNK] [UNK] la [UNK] horror picture show which was really very funny other movies like [UNK] [UNK] from [UNK] space [UNK] to be funny while apparently trying to be serious as to the acting it looks like they [UNK] up [UNK] [UNK] and [UNK] them to [UNK] the whole production [UNK] to them tom [UNK] looks like [UNK] [UNK] there was one decent [UNK] that of the older [UNK] [UNK] on the [UNK]                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               

Original:  b"I saw this film in the worst possible circumstance. I'd already missed 15 minutes when I woke up to it on an international flight between Sydney and Seoul. I didn't know what I was watching, I thought maybe it was a movie of the week, but quickly became riveted by the performance of the lead actress playing a young woman who's child had been kidnapped. The premise started taking twist and turns I didn't see coming and by the end credits I was scrambling through the the in-flight guide to figure out what I had just watched. Turns out I was belatedly discovering Do-yeon Jeon who'd won Best Actress at Cannes for the role. I don't know if Secret Sunshine is typical of Korean cinema but I'm off to the DVD store to discover more."
Round-trip:  i saw this film in the worst possible [UNK] id already [UNK] [UNK] minutes when i [UNK] up to it on an [UNK] [UNK] between [UNK] and [UNK] i didnt know what i was watching i thought maybe it was a movie of the [UNK] but quickly became [UNK] by the performance of the lead actress playing a young woman whos child had been [UNK] the premise started taking twist and turns i didnt see coming and by the end credits i was [UNK] through the the [UNK] [UNK] to figure out what i had just watched turns out i was [UNK] [UNK] [UNK] [UNK] [UNK] [UNK] best actress at [UNK] for the role i dont know if secret [UNK] is typical of [UNK] cinema but im off to the dvd [UNK] to [UNK] more                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     

Original:  b"Hello. I am Paul Raddick, a.k.a. Panic Attack of WTAF, Channel 29 in Philadelphia. Let me tell you about this god awful movie that powered on Adam Sandler's film career but was digitized after a short time.<br /><br />Going Overboard is about an aspiring comedian played by Sandler who gets a job on a cruise ship and fails...or so I thought. Sandler encounters babes that like History of the World Part 1 and Rebound. The babes were supposed to be engaged, but, actually, they get executed by Sawtooth, the meanest cannibal the world has ever known. Adam Sandler fared bad in Going Overboard, but fared better in Big Daddy, Billy Madison, and Jen Leone's favorite, 50 First Dates. Man, Drew Barrymore was one hot chick. Spanglish is red hot, Going Overboard ain't Dooley squat! End of file."
Round-trip:  [UNK] i am paul [UNK] [UNK] [UNK] [UNK] of [UNK] [UNK] [UNK] in [UNK] let me tell you about this god awful movie that [UNK] on [UNK] [UNK] film career but was [UNK] after a short [UNK] br going [UNK] is about an [UNK] [UNK] played by [UNK] who gets a job on a [UNK] [UNK] and [UNK] so i thought [UNK] [UNK] [UNK] that like history of the world part 1 and [UNK] the [UNK] were supposed to be [UNK] but actually they get [UNK] by [UNK] the [UNK] [UNK] the world has ever known [UNK] [UNK] [UNK] bad in going [UNK] but [UNK] better in big [UNK] [UNK] [UNK] and [UNK] [UNK] favorite [UNK] first [UNK] man [UNK] [UNK] was one hot [UNK] [UNK] is red hot going [UNK] [UNK] [UNK] [UNK] end of [UNK]

مدل را ایجاد کنید

ترسیمی از جریان اطلاعات در مدل

در بالا یک نمودار از مدل است.

  1. این مدل می تواند به عنوان یک ساخت tf.keras.Sequential .

  2. اولین لایه است encoder ، که در آن متن به دنباله ای از شاخص های رمز تبدیل می کند.

  3. بعد از رمزگذار یک لایه تعبیه شده است. یک لایه جاسازی یک بردار در هر کلمه ذخیره می کند. هنگام فراخوانی، دنباله های شاخص های کلمه را به دنباله ای از بردارها تبدیل می کند. این بردارها قابل آموزش هستند. پس از آموزش (بر روی داده های کافی)، کلمات با معانی مشابه اغلب بردارهای مشابهی دارند.

    این شاخص مراجعه است بسیار کارآمد تر از عملیات معادل از عبور از یک بردار کد گذاری یک گرم از طریق یک tf.keras.layers.Dense لایه.

  4. یک شبکه عصبی بازگشتی (RNN) ورودی دنباله را با تکرار در بین عناصر پردازش می کند. RNN ها خروجی ها را از یک مرحله زمانی به ورودی خود در مرحله بعدی ارسال می کنند.

    tf.keras.layers.Bidirectional لفاف بسته بندی همچنین می تواند با یک لایه RNN استفاده می شود. این ورودی را از طریق لایه RNN به جلو و عقب منتشر می کند و سپس خروجی نهایی را به هم متصل می کند.

    • مزیت اصلی RNN دو طرفه این است که سیگنال از ابتدای ورودی نیازی به پردازش در تمام مراحل زمانی ندارد تا بر خروجی تأثیر بگذارد.

    • نقطه ضعف اصلی RNN دو طرفه این است که نمی توانید پیش بینی ها را به طور کارآمد پخش کنید زیرا کلمات به انتها اضافه می شوند.

  5. پس از RNN دنباله به یک بردار واحد، تبدیل نموده است دو layers.Dense انجام برخی پردازش نهایی، و تبدیل از این نمایندگی بردار به یک لوجیت تنها به عنوان خروجی طبقه بندی.

کد پیاده سازی آن در زیر است:

model = tf.keras.Sequential([
    encoder,
    tf.keras.layers.Embedding(
        input_dim=len(encoder.get_vocabulary()),
        output_dim=64,
        # Use masking to handle the variable sequence lengths
        mask_zero=True),
    tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(64)),
    tf.keras.layers.Dense(64, activation='relu'),
    tf.keras.layers.Dense(1)
])

لطفاً توجه داشته باشید که مدل ترتیبی Keras در اینجا استفاده می‌شود زیرا همه لایه‌های مدل فقط ورودی واحد دارند و خروجی واحد تولید می‌کنند. در صورتی که می خواهید از لایه RNN حالت دار استفاده کنید، ممکن است بخواهید مدل خود را با API عملکردی Keras یا زیر طبقه بندی مدل بسازید تا بتوانید حالت های لایه RNN را بازیابی و دوباره استفاده کنید. لطفا بررسی کنید راهنمای Keras RNN برای جزئیات بیشتر.

لایه تعبیه استفاده پوشش برای رسیدگی به مختلف توالی طول. همه این لایه ها پس از Embedding پوشش و پشتیبانی:

print([layer.supports_masking for layer in model.layers])
[False, True, True, True, True]

برای تأیید اینکه این کار همانطور که انتظار می رود کار می کند، یک جمله را دو بار ارزیابی کنید. اول، به تنهایی، بنابراین هیچ بالشتکی برای ماسک وجود ندارد:

# predict on a sample text without padding.

sample_text = ('The movie was cool. The animation and the graphics '
               'were out of this world. I would recommend this movie.')
predictions = model.predict(np.array([sample_text]))
print(predictions[0])
[-0.00012211]

اکنون، آن را دوباره در یک دسته با یک جمله طولانی تر ارزیابی کنید. نتیجه باید یکسان باشد:

# predict on a sample text with padding

padding = "the " * 2000
predictions = model.predict(np.array([sample_text, padding]))
print(predictions[0])
[-0.00012211]

برای پیکربندی فرآیند آموزش، مدل Keras را کامپایل کنید:

model.compile(loss=tf.keras.losses.BinaryCrossentropy(from_logits=True),
              optimizer=tf.keras.optimizers.Adam(1e-4),
              metrics=['accuracy'])

مدل را آموزش دهید

history = model.fit(train_dataset, epochs=10,
                    validation_data=test_dataset,
                    validation_steps=30)
Epoch 1/10
391/391 [==============================] - 39s 84ms/step - loss: 0.6454 - accuracy: 0.5630 - val_loss: 0.4888 - val_accuracy: 0.7568
Epoch 2/10
391/391 [==============================] - 30s 75ms/step - loss: 0.3925 - accuracy: 0.8200 - val_loss: 0.3663 - val_accuracy: 0.8464
Epoch 3/10
391/391 [==============================] - 30s 75ms/step - loss: 0.3319 - accuracy: 0.8525 - val_loss: 0.3402 - val_accuracy: 0.8385
Epoch 4/10
391/391 [==============================] - 30s 75ms/step - loss: 0.3183 - accuracy: 0.8616 - val_loss: 0.3289 - val_accuracy: 0.8438
Epoch 5/10
391/391 [==============================] - 30s 75ms/step - loss: 0.3088 - accuracy: 0.8656 - val_loss: 0.3254 - val_accuracy: 0.8646
Epoch 6/10
391/391 [==============================] - 32s 81ms/step - loss: 0.3043 - accuracy: 0.8686 - val_loss: 0.3242 - val_accuracy: 0.8521
Epoch 7/10
391/391 [==============================] - 30s 76ms/step - loss: 0.3019 - accuracy: 0.8696 - val_loss: 0.3315 - val_accuracy: 0.8609
Epoch 8/10
391/391 [==============================] - 32s 76ms/step - loss: 0.3007 - accuracy: 0.8688 - val_loss: 0.3245 - val_accuracy: 0.8609
Epoch 9/10
391/391 [==============================] - 31s 77ms/step - loss: 0.2981 - accuracy: 0.8707 - val_loss: 0.3294 - val_accuracy: 0.8599
Epoch 10/10
391/391 [==============================] - 31s 78ms/step - loss: 0.2969 - accuracy: 0.8742 - val_loss: 0.3218 - val_accuracy: 0.8547
test_loss, test_acc = model.evaluate(test_dataset)

print('Test Loss:', test_loss)
print('Test Accuracy:', test_acc)
391/391 [==============================] - 15s 38ms/step - loss: 0.3185 - accuracy: 0.8582
Test Loss: 0.3184521794319153
Test Accuracy: 0.8581600189208984
plt.figure(figsize=(16, 8))
plt.subplot(1, 2, 1)
plot_graphs(history, 'accuracy')
plt.ylim(None, 1)
plt.subplot(1, 2, 2)
plot_graphs(history, 'loss')
plt.ylim(0, None)
(0.0, 0.6627909764647484)

png

پیش بینی یک جمله جدید را اجرا کنید:

اگر پیش بینی >= 0.0 باشد، مثبت است وگرنه منفی است.

sample_text = ('The movie was cool. The animation and the graphics '
               'were out of this world. I would recommend this movie.')
predictions = model.predict(np.array([sample_text]))

دو یا چند لایه LSTM را روی هم قرار دهید

Keras لایه مکرر دو حالت در دسترس است که توسط کنترل return_sequences استدلال سازنده:

  • اگر False آن را تنها در آخرین خروجی برای هر توالی ورودی را برمی گرداند (یک تانسور 2D شکل (batch_size، output_features)). این پیش فرض است که در مدل قبلی استفاده شده است.

  • اگر True توالی کامل از خروجی های پی در پی برای هر timestep بازگردانده می شود (یک تانسور 3D شکل (batch_size, timesteps, output_features) ).

اینجا چیزی است که به نظر می رسد جریان اطلاعات مانند با return_sequences=True :

لایه بندی شده_دو جهته

نکته جالب در مورد استفاده از RNN با return_sequences=True است که خروجی هنوز دارای 3 محور، مانند ورودی، بنابراین می توان آن را به یکی دیگر از لایه RNN گذشت، مثل این:

model = tf.keras.Sequential([
    encoder,
    tf.keras.layers.Embedding(len(encoder.get_vocabulary()), 64, mask_zero=True),
    tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(64,  return_sequences=True)),
    tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(32)),
    tf.keras.layers.Dense(64, activation='relu'),
    tf.keras.layers.Dropout(0.5),
    tf.keras.layers.Dense(1)
])
model.compile(loss=tf.keras.losses.BinaryCrossentropy(from_logits=True),
              optimizer=tf.keras.optimizers.Adam(1e-4),
              metrics=['accuracy'])
history = model.fit(train_dataset, epochs=10,
                    validation_data=test_dataset,
                    validation_steps=30)
Epoch 1/10
391/391 [==============================] - 71s 149ms/step - loss: 0.6502 - accuracy: 0.5625 - val_loss: 0.4923 - val_accuracy: 0.7573
Epoch 2/10
391/391 [==============================] - 55s 138ms/step - loss: 0.4067 - accuracy: 0.8198 - val_loss: 0.3727 - val_accuracy: 0.8271
Epoch 3/10
391/391 [==============================] - 54s 136ms/step - loss: 0.3417 - accuracy: 0.8543 - val_loss: 0.3343 - val_accuracy: 0.8510
Epoch 4/10
391/391 [==============================] - 53s 134ms/step - loss: 0.3242 - accuracy: 0.8607 - val_loss: 0.3268 - val_accuracy: 0.8568
Epoch 5/10
391/391 [==============================] - 53s 135ms/step - loss: 0.3174 - accuracy: 0.8652 - val_loss: 0.3213 - val_accuracy: 0.8516
Epoch 6/10
391/391 [==============================] - 52s 132ms/step - loss: 0.3098 - accuracy: 0.8671 - val_loss: 0.3294 - val_accuracy: 0.8547
Epoch 7/10
391/391 [==============================] - 53s 134ms/step - loss: 0.3063 - accuracy: 0.8697 - val_loss: 0.3158 - val_accuracy: 0.8594
Epoch 8/10
391/391 [==============================] - 52s 132ms/step - loss: 0.3043 - accuracy: 0.8692 - val_loss: 0.3184 - val_accuracy: 0.8521
Epoch 9/10
391/391 [==============================] - 53s 133ms/step - loss: 0.3016 - accuracy: 0.8704 - val_loss: 0.3208 - val_accuracy: 0.8609
Epoch 10/10
391/391 [==============================] - 54s 136ms/step - loss: 0.2975 - accuracy: 0.8740 - val_loss: 0.3301 - val_accuracy: 0.8651
test_loss, test_acc = model.evaluate(test_dataset)

print('Test Loss:', test_loss)
print('Test Accuracy:', test_acc)
391/391 [==============================] - 26s 65ms/step - loss: 0.3293 - accuracy: 0.8646
Test Loss: 0.329334557056427
Test Accuracy: 0.8646399974822998
# predict on a sample text without padding.

sample_text = ('The movie was not good. The animation and the graphics '
               'were terrible. I would not recommend this movie.')
predictions = model.predict(np.array([sample_text]))
print(predictions)
[[-1.6796288]]
plt.figure(figsize=(16, 6))
plt.subplot(1, 2, 1)
plot_graphs(history, 'accuracy')
plt.subplot(1, 2, 2)
plot_graphs(history, 'loss')

png

اتمام دیگر لایه های مکرر موجود مانند لایه GRU .

اگر شما در حال ساخت و ساز در RNNs سفارشی interestied، را ببینید راهنمای RNN Keras .