مؤتمر Google I / O هو التفاف! تابع جلسات TensorFlow اعرض الجلسات

تصنيف النص مع RNN

عرض على TensorFlow.org تشغيل في Google Colab عرض المصدر على جيثب تحميل دفتر

هذا البرنامج التعليمي تصنيف النص القطارات على الشبكة العصبية المتكررة على IMDB فيلم كبير مراجعة بيانات لتحليل المشاعر.

يثبت

import numpy as np

import tensorflow_datasets as tfds
import tensorflow as tf

tfds.disable_progress_bar()

استيراد matplotlib وإنشاء وظيفة مساعد لمؤامرة الرسوم البيانية:

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])

إعداد خط أنابيب الإدخال

وIMDB كبير فيلم استعراض مجموعة البيانات هو ثنائي تصنيف مجموعة البيانات للجميع الاستعراضات إما إيجابية أو المشاعر السلبية.

تحميل بيانات باستخدام TFDS . انظر تحميل النص تعليمي للحصول على تفاصيل حول كيفية تحميل هذا النوع من البيانات يدويا.

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))

يؤدي هذا في البداية إلى إرجاع مجموعة بيانات (نص ، أزواج ملصقات):

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

التالي خلط البيانات للتدريب وإنشاء دفعات من هذه (text, label) أزواج:

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]

قم بإنشاء برنامج تشفير النص

النص الخام تحميلها من قبل tfds يحتاج إلى معالجتها قبل أنها يمكن أن تستخدم في نموذج. إن أبسط طريقة لنص عملية التدريب يستخدم TextVectorization طبقة. تتمتع هذه الطبقة بالعديد من الإمكانيات ، لكن هذا البرنامج التعليمي يتمسك بالسلوك الافتراضي.

خلق طبقة، وتمرير نص ورقة العمل إلى الطبقة .adapt الأسلوب:

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

و .adapt طريقة يحدد المفردات الطبقة. فيما يلي أول 20 رمزًا. بعد الحشو والرموز غير المعروفة يتم فرزها حسب التردد:

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')

بمجرد تعيين المفردات ، يمكن للطبقة ترميز النص إلى مؤشرات. و0 مبطن، والتنسورات مؤشرات لأطول تسلسل في الدفعة (إلا إذا قمت بتعيين 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]])

مع الإعدادات الافتراضية ، لا يمكن عكس العملية تمامًا. هناك ثلاثة أسباب رئيسية لذلك:

  1. القيمة الافتراضية ل preprocessing.TextVectorization الصورة standardize الحجة "lower_and_strip_punctuation" .
  2. ينتج عن حجم المفردات المحدود وعدم وجود احتياطي قائم على الأحرف بعض الرموز المميزة غير المعروفة.
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]

قم بإنشاء النموذج

رسم لتدفق المعلومات في النموذج

أعلاه هو رسم تخطيطي للنموذج.

  1. هذا النموذج يمكن أن يكون بناء على أنه tf.keras.Sequential .

  2. الطبقة الأولى هي encoder ، والذي يحول النص إلى سلسلة من المؤشرات رمزية.

  3. بعد التشفير طبقة التضمين. تخزن طبقة التضمين متجهًا واحدًا لكل كلمة. عندما يتم استدعاؤه ، فإنه يحول متواليات فهارس الكلمات إلى متواليات من المتجهات. هذه النواقل قابلة للتدريب. بعد التدريب (على بيانات كافية) ، غالبًا ما يكون للكلمات ذات المعاني المتشابهة نواقل متشابهة.

    هذا مؤشر البحث هو أكثر كفاءة بكثير من عملية ما يعادل تمرير ناقلات المشفرة الساخنة واحد من خلال tf.keras.layers.Dense طبقة.

  4. تعالج الشبكة العصبية المتكررة (RNN) إدخال التسلسل عن طريق التكرار خلال العناصر. تمرر RNN المخرجات من خطوة زمنية إلى إدخالها في الخطوة الزمنية التالية.

    و tf.keras.layers.Bidirectional يمكن أيضا أن تستخدم المجمع مع طبقة RNN. يؤدي هذا إلى نشر المدخلات للأمام وللخلف من خلال طبقة RNN ثم تسلسل الإخراج النهائي.

    • الميزة الرئيسية لـ RNN ثنائية الاتجاه هي أن الإشارة من بداية الإدخال لا تحتاج إلى المعالجة طوال الطريق خلال كل خطوة زمنية للتأثير على المخرجات.

    • العيب الرئيسي لـ RNN ثنائي الاتجاه هو أنه لا يمكنك دفق التنبؤات بكفاءة حيث تتم إضافة الكلمات إلى النهاية.

  5. بعد RNN حولت تسلسل إلى ناقل واحد واثنين layers.Dense القيام ببعض المعالجة النهائية، وتحويل من هذا التمثيل متجه إلى اللوغاريتمي واحد والإخراج التصنيف.

الكود المطلوب تنفيذه أدناه:

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)
])

يرجى ملاحظة أنه يتم استخدام نموذج Keras المتسلسل هنا نظرًا لأن جميع الطبقات في النموذج لها مدخل واحد فقط وتنتج مخرجات فردية. إذا كنت ترغب في استخدام طبقة RNN ذات الحالة ، فقد ترغب في بناء النموذج الخاص بك باستخدام واجهة برمجة تطبيقات Keras الوظيفية أو فئة فرعية نموذجية بحيث يمكنك استرداد حالات طبقة RNN وإعادة استخدامها. يرجى مراجعة دليل Keras RNN لمزيد من التفاصيل.

طبقة تضمين استخدامات اخفاء للتعامل مع متفاوتة تسلسل أطوال. جميع طبقات بعد Embedding اخفاء الدعم:

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

للتأكد من أن هذا يعمل كما هو متوقع ، قم بتقييم الجملة مرتين. أولاً ، بمفردك لذلك لا توجد حشوة لإخفائها:

# 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]

الآن ، قم بتقييمه مرة أخرى في مجموعة بجملة أطول. يجب أن تكون النتيجة متطابقة:

# predict on a sample text with padding

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

قم بتجميع نموذج Keras لتكوين عملية التدريب:

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

تدريب النموذج

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)

بي إن جي

قم بتشغيل توقع على جملة جديدة:

إذا كان التوقع> = 0.0 ، فهو إيجابي ، وإلا سيكون سالبًا.

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]))

قم بتكديس طبقتين أو أكثر من طبقات LSTM

Keras طبقات المتكررة واثنين من وسائط المتاحة التي تسيطر عليها return_sequences حجة المنشئ:

  • إذا False فإنه يعود فقط إخراج الأخير لكل تسلسل المدخلات (أ موتر 2D الشكل (batch_size، output_features)). هذا هو الإعداد الافتراضي المستخدم في النموذج السابق.

  • إذا True يتم إرجاع تسلسل كامل النواتج المتعاقبة لكل خطوة زمنية (أ موتر 3D الشكل (batch_size, timesteps, output_features) ).

هنا هو ما تدفق تبدو المعلومات مثل مع return_sequences=True :

متعدد الاتجاهات

والشيء المثير للاهتمام حول استخدام RNN مع return_sequences=True هو ان الانتاج لا يزال لديه 3 محاور، مثل المدخلات، لذلك يمكن أن تنتقل إلى طبقة RNN آخر، مثل هذا:

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')

بي إن جي

تحقق من طبقات المتكررة القائمة الأخرى مثل طبقات GRU .

إذا كنت interestied في بناء RNNs مخصصة، راجع دليل RNN Keras .