zhengmingzhang / Deep-RNN

Sentence Generation using PyTorch

ds-ga-1008-a2

NYU course Spring 2017 Assignment 2

This sample code is from https://github.com/pytorch/examples/tree/master/word_language_model.

Word-level language modeling RNN

This example trains a multi-layer RNN (Elman, GRU, or LSTM) on a language modeling task. By default, the training script uses the PTB dataset, provided. The trained model can then be used by the generate script to generate new text.

python main.py --epochs 6        # Train a LSTM on PTB, reaching perplexity of 117.61
python main.py --epochs 6 --tied # Train a tied LSTM on PTB, reaching perplexity of 110.44
python main.py --tied            # Train a tied LSTM on PTB for 40 epochs, reaching perplexity of 87.17
python generate.py                      # Generate samples from the trained LSTM model.

The model uses the nn.RNN module (and its sister modules nn.GRU and nn.LSTM)

During training, if a keyboard interrupt (Ctrl-C) is received, training is stopped and the current model is evaluted against the test dataset.

The main.py script accepts the following arguments:

optional arguments:
  -h, --help         show this help message and exit
  --data DATA        location of the data corpus
  --model MODEL      type of recurrent net (RNN_TANH, RNN_RELU, LSTM, GRU)
  --emsize EMSIZE    size of word embeddings
  --nhid NHID        humber of hidden units per layer
  --nlayers NLAYERS  number of layers
  --lr LR            initial learning rate
  --clip CLIP        gradient clipping
  --epochs EPOCHS    upper epoch limit
  --batch-size N     batch size
  --bptt BPTT        sequence length
  --dropout DROPOUT  dropout applied to layers (0 = no dropout)
  --decay DECAY      learning rate decay per epoch
  --tied             tie the word embedding and softmax weights
  --seed SEED        random seed
  --log-interval N   report interval
  --optim            type of optimizer
  --tsnedim          size of the subset needed for the tsne
  --save SAVE        path to save the final model

With these arguments, a variety of models can be tested. As an example, the following arguments produce slower but better models:

python main.py --emsize 650 --nhid 650 --dropout 0.5 --epochs 40           # Test perplexity of 80.97
python main.py --emsize 650 --nhid 650 --dropout 0.5 --epochs 40 --tied    # Test perplexity of 75.96
python main.py --emsize 1500 --nhid 1500 --dropout 0.65 --epochs 40        # Test perplexity of 77.42
python main.py --emsize 1500 --nhid 1500 --dropout 0.65 --epochs 40 --tied # Test perplexity of 72.30

These perplexities are equal or better than Recurrent Neural Network Regularization (Zaremba et al. 2014) and are similar to Using the Output Embedding to Improve Language Models (Press & Wolf 2016 and Tying Word Vectors and Word Classifiers: A Loss Framework for Language Modeling (Inan et al. 2016), though both of these papers have improved perplexities by using a form of recurrent dropout (variational dropout).