Create your own GitHub profile
Sign up for your own profile on GitHub, the best place to host code, manage projects, and build software alongside 50 million developers.
Sign up
-
Hugging Face
- Chinatown, New York
- Sign in to view email
- Lysand.re
Pinned
1,683 contributions in the last year
Activity overview
Contributed to
huggingface/transformers,
ProjectBlackFalcon/Roks,
huggingface/swift-coreml-transformers
and 5 other
repositories
Contribution activity
May 2020
Created a pull request in huggingface/transformers that received 2 comments
- [Docs, Notebook] Include generation pipeline
- Question Answering for TF trainer
- [Marian Fixes] prevent predicting pad_token_id before softmax, support language codes, name multilingual models
- added functionality for electra classification head
- Simplify cache vars and allow for TRANSFORMERS_CACHE env
- Allow gpt2 to be exported to valid ONNX
- Conversion script to export transformers models to ONNX IR.
- [tests] make torch pipelines tests faster with smaller models
- Add migrating from `pytorch-transformers`
- Align sentiment-analysis' tokenizer (currently cased) to the model (uncased)
- Tokenizer.decode_batch convenience method
- Add ALBERT to the Tensorflow to Pytorch model conversion cli
- [Pipeline, Generation] tf generation pipeline bug
- Remove hard-coded pad token id in distilbert and albert
- BIG Reorganize examples
- Ensure fast tokenizer can construct single-element tensor without pad token
- change order pytorch/tf in readme
- TF version of the trainer
- Include ElectraPreTrainedModel into __init__
- Tpu trainer
- Rewritten batch support in pipelines.
- [Marian] documentation and AutoModel support
- FIXME(Actually test multi input pipelines)
- Reformer
- Proposal: saner num_labels in configs.
- Some pull request reviews not shown.
5
contributions
in private repositories
May 6 – May 13