文本分类，RNN模型

Apply the dropout technique in Gal and Ghahramani, "Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning" (ht

The composition layer.

Implement ESIM model using the modules defined above.

Implement embedding layer.

BiLSTM encoder which encodes both the premise and hypothesis.

Inference composition described in paper section 3.3

Implement the final linear layer.

The local inference model introduced in the paper.

Apply maxing pooling and average pooling to the outputs of LSTM.

Dropout layer for the inputs of RNNs. Apply the same dropout mask to all the elements of the same sequence in a batch of sequences of siz

Attention layer taking premises and hypotheses encoded by an RNN as input and computing the soft attention between their elements. The do