Texar-Pytorch v0.1.4

• Add tests for python3.8 and python3.9 #340
• workflow and code update for numpy versions from 1.15 to 1.21. #352
• Move out HParams and SpecialTokens to asyml-utilities #353
• Several bug fixes #335 #345 #351

Texar-PyTorch v0.1.2

New features

1. Integrated texar-pytorch with NNI and AdaptDL for distributed adaptive API (#331)
2. Integrated with NNI for hyperparameter tuning (#324)
3. Add Information Loss based on KL divergence (#328)
4. Allow WordpieceTokenizer to return the original character spans (#332)
5. Add a few modules: RNN Classifier (#303), SpanBERT (#300)

Feature improvements:

1. Fix a few documentation issues.

Fixes

1. Fix a type error in Beam Search top-k index (#330)
2. Fix a file operation error in Executor (#323)
3. Fix some evaluator bugs and related file handling (#320)
4. Fix a problem where metrics cannot be pickled (#319)

Texar-PyTorch v0.1.1

New features

• Support PyTorch 1.3. (#249)
• Add T5 modules (T5Encoder, T5Decoder, and T5EncoderDecoder). (#280)
• Add T5Tokenizer. (#283)
• Support PyTorch 1.4. (#291)

Feature improvements

• Refactor the interface of GPT2 modules. (#238)
• Support gpt2-xl checkpoint file in GPT2 modules. (#242)
• Add code coverage check in CI. (#245)
• Update the vocabulary files of RoBERTa modules. (#255)
• Disable codecov/patch check in CI. (#265)
• Provide option to freeze the embedding parameters. (#271)
• Add encode_text_for_generation function in XLNetTokenizer. (#278)
• Use warning instead of error in map_token_to_id function. (#285)
• Add copyright header to unit tests. (#287)
• Remove duplicated pytest in CI. (#289)
• Update the versions of pylint, flake8, and mypy in CI. (#292)

Fixes

• Fix the documentation issues in SentencePieceTokenizer. (#236)
• Fix the bugs in RoBERTa checkpoint file loading procedure. (#241)
• Fix the documentation issues in Executor. (#244)
• Fix the documentation issues in gpt-2 example. (#250)
• Fix the bugs in bidirectional_dynamic_rnn and dynamic_rnn functions. (#252)
• Fix the bugs in vae_text example. (#253)
• Fix the bugs in sentence_classifier example. (#262)
• Fix the path error when installing texar-pytorch in Windows. (#268)
• Fix the bugs in XLNetTokenizer. (#273)
• Fix the bugs in download_checkpoint function. (#274)
• Fix the bugs in google drive downloading function. (#275)
• Fix the bugs in the unit test of GPT2Decoder. (#288)
• Fix the documentation issues in Decoder module. (#290)

Texar-PyTorch v0.1.0

First formal release of Texar-PyTorch

Texar-PyTorch v0.0.1

Initial release