sgugger Follow

Sylvain Gugger sgugger

Research Engineer at HuggingFace

913 followers · 3 following · 12

Highlights

Arctic Code Vault Contributor

Organizations

Popular repositories

Deep-Learning

A few notebooks about deep learning in pytorch

Jupyter Notebook 473 98
Adam-experiments

Experiments with Adam/AdamW/amsgrad

Python 164 33
SwiftData

Swift 12 1
fastai

Forked from fastai/fastai

The fast.ai deep learning library, lessons, and tutorials

Jupyter Notebook 3 1
transformer_train

Python 3
swift-training

A generic training loop for S4TF

Swift 3 1

3,297 contributions in the last year

Contribution activity

October 2020

Created a pull request in huggingface/transformers that received 2 comments

Oct 19

New run glue script

What does this PR do? This PR cleans up the run_glue.py script to use the Datasets library. Along the way it adds a few fixes in Trainer. The scrip…

+329 −168 • 2 comments

Doc styling Oct 23
Reload checkpoint Oct 22
Only log total_flos at the end of training Oct 22
Expose the Flax code quality problems Oct 19
Add Flax dummy objects Oct 19
Small fixes to HP search Oct 15
Small fixes to NotebookProgressCallback Oct 15
Add specific notebook ProgressCalback Oct 14
Don't use `store_xxx` on optional bools Oct 14
Add predict step accumulation Oct 13
Fix typo in all model docs Oct 11
Fix flaky test in test_trainer Oct 9
Fix title level in Blenderbot doc Oct 9
Better links for models in README and doc index Oct 9
Trainer callbacks Oct 5
Documentation fixes Oct 5
Expand test to locate flakiness Oct 5
Remove labels from the RagModel example Oct 3
Allow nested tensors in predicted logits Oct 2
Allow soft dependencies in the namespace with ImportErrors at use Oct 2
Check and update model list in index.rst automatically Oct 1
Fix post_init of some TrainingArguments Oct 1
Cleanup documentation for BART, Marian, MBART and Pegasus Oct 1
Fix seq2seq example test Oct 1

Fix GLUE metric description Oct 14

Add mixed precision evaluation Oct 26
[cleanup] pegasus,marian,mbart pytorch tests Oct 25
[tokenizers] Fixing #8001 - Adding tests on tokenizers serialization Oct 23
[tests|tokenizers] Refactoring pipelines test backbone - Small tokenizers improvements - General tests speedups Oct 23
FillMaskPipeline: support passing top_k on __call__ Oct 22
support relative path for best_model_checkpoint Oct 22
fix 'encode_plus' docstring for 'special_tokens_mask' (0s and 1s were reversed) Oct 21
Move NoLayerEmbedTokens Oct 21
TensorBoard/Wandb/optuna/raytune integration improvements. Oct 21
Fix comet_ml import and add ensure availability Oct 21
[ProphetNet] Correct Doc string example Oct 21
[testing] slow tests should be marked as slow Oct 21
[PretrainedConfig] Fix save pretrained config for edge case Oct 21
[multiple models] skip saving/loading deterministic state_dict keys Oct 21
# Add whole word mask support for lm fine-tune Oct 20
Respect the 119 line chars Oct 20
[testing] experiment with a different way of skipping torch-only test modules Oct 20
[flax] fix repo_check Oct 20
New run glue script Oct 19
Trainer with Iterable Dataset Oct 19
labels and decoder_input_ids to Glossary Oct 19
Add TFBartForConditionalGeneration Oct 19
[CIs] report slow tests add --durations=0 to some pytest jobs Oct 19
ProphetNet Oct 17
[testing] remove USE_CUDA Oct 17
Some pull request reviews not shown.

Sep	OCT	Nov
	26
2019	2020	2021

Sylvain Gugger sgugger

Block or report sgugger

Highlights

Organizations

Create your own GitHub profile

Popular repositories

3,297 contributions in the last year

Contribution activity

October 2020

Created a pull request in huggingface/transformers that received 2 comments

New run glue script

Essential cookies

Always active

Analytics cookies