# How to build a simple text classifier with TF-Hub

TF-Hub is a platform to share machine learning expertise packaged in reusable resources, notably pre-trained modules. This tutorial is organized into two main parts.

Introduction: Training a text classifier with TF-Hub

We will use a TF-Hub text embedding module to train a simple sentiment classifier with a reasonable baseline accuracy. We will then analyze the predictions to make sure our model is reasonable and propose improvements to increase the accuracy.

In this section, we will use various TF-Hub modules to compare their effect on the accuracy of the estimator and demonstrate advantages and pitfalls of transfer learning.

## Preparing the enviroment

# Install the latest Tensorflow version.
# Install TF-Hub.
!pip install -q tensorflow-hub


More detailed information about installing Tensorflow can be found at https://www.tensorflow.org/install/.

import tensorflow as tf
import tensorflow_hub as hub
import matplotlib.pyplot as plt
import numpy as np
import os
import pandas as pd
import re
import seaborn as sns

    /usr/local/lib/python2.7/dist-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from float to np.floating is deprecated. In future, it will be treated as np.float64 == np.dtype(float).type.
from ._conv import register_converters as _register_converters


# Getting started

## Data

We will try to solve the Large Movie Review Dataset v1.0 task from Mass et al. The dataset consists of IMDB movie reviews labeled by positivity from 1 to 10. The task is to label the reviews as negative or positive.

# Load all files from a directory in a DataFrame.
data = {}
data["sentence"] = []
data["sentiment"] = []
for file_path in os.listdir(directory):
with tf.gfile.GFile(os.path.join(directory, file_path), "r") as f:
data["sentiment"].append(re.match("\d+_(\d+)\.txt", file_path).group(1))
return pd.DataFrame.from_dict(data)

# Merge positive and negative examples, add a polarity column and shuffle.
pos_df["polarity"] = 1
neg_df["polarity"] = 0
return pd.concat([pos_df, neg_df]).sample(frac=1).reset_index(drop=True)

dataset = tf.keras.utils.get_file(
fname="aclImdb.tar.gz",
origin="http://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz",
extract=True)

"aclImdb", "train"))
"aclImdb", "test"))

return train_df, test_df

# Reduce logging output.
tf.logging.set_verbosity(tf.logging.ERROR)


    Downloading data from http://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz
84131840/84125825 [==============================] - 1s 0us/step
84140032/84125825 [==============================] - 1s 0us/step

sentence sentiment polarity
0 I just rented this today....heard lots of good... 1 0
1 Outrage is pretty good movie! Robert Culp was ... 10 1
2 OK, as everyone has pointed out, this film is ... 3 0
3 I am a current A.S.L. Student & was forced to ... 4 0
4 Redundant, but again the case. If you enjoy th... 2 0

## Model

### Input functions

Estimator framework provides input functions that wrap Pandas dataframes.

# Training input on the whole training set with no limit on training epochs.
train_input_fn = tf.estimator.inputs.pandas_input_fn(
train_df, train_df["polarity"], num_epochs=None, shuffle=True)

# Prediction on the whole training set.
predict_train_input_fn = tf.estimator.inputs.pandas_input_fn(
train_df, train_df["polarity"], shuffle=False)
# Prediction on the test set.
predict_test_input_fn = tf.estimator.inputs.pandas_input_fn(
test_df, test_df["polarity"], shuffle=False)


### Feature columns

TF-Hub provides a feature column that applies a module on the given text feature and passes further the outputs of the module. In this tutorial we will be using the nnlm-en-dim128 module. For the purpose of this tutorial, the most important facts are:

• The module takes a batch of sentences in a 1-D tensor of strings as input.
• The module is responsible for preprocessing of sentences (e.g. removal of punctuation and splitting on spaces).
• The module works with any input (e.g. nnlm-en-dim128 hashes words not present in vocabulary into ~20.000 buckets).
embedded_text_feature_column = hub.text_embedding_column(
key="sentence",


### Estimator

For classification we can use a DNN Classifier (note further remarks about different modelling of the label function at the end of the tutorial).

estimator = tf.estimator.DNNClassifier(
hidden_units=[500, 100],
feature_columns=[embedded_text_feature_column],
n_classes=2,


### Training

Train the estimator for a reasonable amount of steps.

# Training for 1,000 steps means 128,000 training examples with the default
# batch size. This is roughly equivalent to 5 epochs since the training dataset
# contains 25,000 examples.
estimator.train(input_fn=train_input_fn, steps=1000);


# Prediction

Run predictions for both training and test set.

train_eval_result = estimator.evaluate(input_fn=predict_train_input_fn)
test_eval_result = estimator.evaluate(input_fn=predict_test_input_fn)

print "Training set accuracy: {accuracy}".format(**train_eval_result)
print "Test set accuracy: {accuracy}".format(**test_eval_result)

    Training set accuracy: 0.802160024643
Test set accuracy: 0.792879998684


## Confusion matrix

We can visually check the confusion matrix to undestand the distribution of misclassifications.

def get_predictions(estimator, input_fn):
return [x["class_ids"][0] for x in estimator.predict(input_fn=input_fn)]

LABELS = [
"negative", "positive"
]

# Create a confusion matrix on training data.
with tf.Graph().as_default():
cm = tf.confusion_matrix(train_df["polarity"],
get_predictions(estimator, predict_train_input_fn))
with tf.Session() as session:
cm_out = session.run(cm)

# Normalize the confusion matrix so that each row sums to 1.
cm_out = cm_out.astype(float) / cm_out.sum(axis=1)[:, np.newaxis]

sns.heatmap(cm_out, annot=True, xticklabels=LABELS, yticklabels=LABELS);
plt.xlabel("Predicted");
plt.ylabel("True");


# Further improvements

1. Regression on sentiment: we used a classifier to assign each example into a polarity class. But we actually have another categorical feature at our disposal - sentiment. Here classes actually represent a scale and the underlying value (positive/negative) could be well mapped into a continuous range. We could make use of this property by computing a regression (DNN Regressor) instead of a classification (DNN Classifier).
2. Larger module: for the purposes of this tutorial we used a small module to restrict the memory use. There are modules with larger vocabularies and larger embedding space that could give additional accuracy points.
3. Parameter tuning: we can improve the accuracy by tuning the meta-parameters like the learning rate or the number of steps, especially if we use a different module. A validation set is very important if we want to get any reasonable results, because it is very easy to set-up a model that learns to predict the training data without generalizing well to the test set.
4. More complex model: we used a module that computes a sentence embedding by embedding each individual word and then combining them with average. One could also use a sequential module (e.g. Universal Sentence Encoder module) to better capture the nature of sentences. Or an ensemble of two or more TF-Hub modules.
5. Regularization: to prevent overfitting, we could try to use an optimizer that does some sort of regularization, for example Proximal Adagrad Optimizer.

Transfer learning makes it possible to save training resources and to achieve good model generalization even when training on a small dataset. In this part, we will demonstrate this by training with two different TF-Hub modules:

• nnlm-en-dim128 - pretrained text embedding module,
• random-nnlm-en-dim128 - text embedding module that has same vocabulary and network as nnlm-en-dim128, but the weights were just randomly initialized and never trained on real data.

And by training in two modes:

• training only the classifier (i.e. freezing the module), and
• training the classifier together with the module.

Let's run a couple of trainings and evaluations to see how using a various modules can affect the accuracy.

def train_and_evaluate_with_module(hub_module, train_module=False):
embedded_text_feature_column = hub.text_embedding_column(
key="sentence", module_spec=hub_module, trainable=train_module)

estimator = tf.estimator.DNNClassifier(
hidden_units=[500, 100],
feature_columns=[embedded_text_feature_column],
n_classes=2,

estimator.train(input_fn=train_input_fn, steps=1000)

train_eval_result = estimator.evaluate(input_fn=predict_train_input_fn)
test_eval_result = estimator.evaluate(input_fn=predict_test_input_fn)

training_set_accuracy = train_eval_result["accuracy"]
test_set_accuracy = test_eval_result["accuracy"]

return {
"Training accuracy": training_set_accuracy,
"Test accuracy": test_set_accuracy
}

results = {}
results["nnlm-en-dim128"] = train_and_evaluate_with_module(
results["nnlm-en-dim128-with-module-training"] = train_and_evaluate_with_module(
results["random-nnlm-en-dim128"] = train_and_evaluate_with_module(
results["random-nnlm-en-dim128-with-module-training"] = train_and_evaluate_with_module(


Let's look at the results.

pd.DataFrame.from_dict(results, orient="index")

Training accuracy Test accuracy
nnlm-en-dim128 0.80176 0.79324
nnlm-en-dim128-with-module-training 0.94912 0.86996
random-nnlm-en-dim128 0.72244 0.67456
random-nnlm-en-dim128-with-module-training 0.76584 0.72180

We can already see some patterns, but first we should establish the baseline accuracy of the test set - the lower bound that can be achieved by outputting only the label of the most represented class:

estimator.evaluate(input_fn=predict_test_input_fn)["accuracy_baseline"]

0.5


Assigning the most represented class will give us accuracy of 50%. There are a couple of things to notice here:

1. Maybe surprisingly, a model can still be learned on top of fixed, random embeddings. The reason is that even if every word in the dictionary is mapped to a random vector, the estimator can separate the space purely using its fully connected layers.
2. Allowing training of the module with random embeddings increases both training and test accuracy as oposed to training just the classifier.
3. Training of the module with pre-trained embeddings also increases both accuracies. Note however the overfitting on the training set. Training a pre-trained module can be dangerous even with regularization in the sense that the embedding weights no longer represent the language model trained on diverse data, instead they converge to the ideal representation of the new dataset.