Attention is all you need: A Pytorch Implementation

This is a PyTorch implementation of the Transformer model in "Attention is All You Need" (Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, Illia Polosukhin, arxiv, 2017).

A novel sequence to sequence framework utilizes the self-attention mechanism, instead of Convolution operation or Recurrent structure, and achieve the state-of-the-art performance on WMT 2014 English-to-German translation task. (2017/06/12)

The official Tensorflow Implementation can be found in: tensorflow/tensor2tensor.

To learn more about self-attention mechanism, you could read "A Structured Self-attentive Sentence Embedding".

The project support training and translation with trained model now.

Note that this project is still a work in progress.

BPE related parts are not yet fully tested.

If there is any suggestion or error, feel free to fire an issue to let me know. :)

Usage
WMT'16 Multimodal Translation: de-en

An example of training for the WMT'16 Multimodal Translation task (http://www.statmt.org/wmt16/multimodal-task.html).

0) Download the spacy language model.

```bash

conda install -c conda-forge spacy

python -m spacy download en
python -m spacy download de
```

1) Preprocess the data with torchtext and spacy.

bash
python preprocess.py -lang_src de -lang_trg en -share_vocab -save_data m30k_deen_shr.pkl

2) Train the model

bash
python train.py -data_pkl m30k_deen_shr.pkl -log m30k_deen_shr -embs_share_weight -proj_share_weight -label_smoothing -output_dir output -b 256 -warmup 128000 -epoch 400

3) Test the model

bash
python translate.py -data_pkl m30k_deen_shr.pkl -model trained.chkpt -output prediction.txt

[(WIP)] WMT'17 Multimodal Translation: de-en w/ BPE
1) Download and preprocess the data with bpe:

Since the interfaces is not unified, you need to switch the main function call from main_wo_bpe to main.

bash
python preprocess.py -raw_dir /tmp/raw_deen -data_dir ./bpe_deen -save_data bpe_vocab.pkl -codes codes.txt -prefix deen

2) Train the model

bash
python train.py -data_pkl ./bpe_deen/bpe_vocab.pkl -train_path ./bpe_deen/deen-train -val_path ./bpe_deen/deen-val -log deen_bpe -embs_share_weight -proj_share_weight -label_smoothing -output_dir output -b 256 -warmup 128000 -epoch 400

3) Test the model (not ready)

• TODO:
  
  
  • Load vocabulary.
  
  
  • Perform decoding after the translation.
  
  
  
  

Performance
Training

• Parameter settings:


• batch size 256


• warmup step 4000


• epoch 200


• lr_mul 0.5


• label smoothing


• do not apply BPE and shared vocabulary


• target embedding / pre-softmax linear layer weight sharing.

Testing
- coming soon.
TODO

• Evaluation on the generated text.


• Attention weight plot.

Acknowledgement

• The byte pair encoding parts are borrowed from subword-nmt.


• The project structure, some scripts and the dataset preprocessing steps are heavily borrowed from OpenNMT/OpenNMT-py.


• Thanks for the suggestions from @srush, @iamalbert, @Zessay, @JulesGM, @ZiJianZhao, and @huanghoujing.