NielsRogge/transformers

Pull Requests Commits

First draft of ImageProcessor

NielsRogge committed 4 years ago

aa99ecf8

Fixed minor spelling mistakes (#10489)

WybeKoper committed 4 years ago

Verified 5dc303e2

Generate can return cross-attention weights too (#10493)

Mehrad0711 committed 4 years ago

Verified 1750e629

Changed `num_beams` to `num_beams // num_beam_groups` when initialising `PrefixConstrainedLogitsProcessor` in `_get_logits_processor` to fix compatibility issue when constrained decoding is used together with grouped beam search (#10475)

mnschmit committed 4 years ago

Verified b0138422

Add I-BERT to README (#10462)

LysandreJik committed 4 years ago

Verified 0c232519

Remove Anthony from the bug reports in Transformers

LysandreJik committed 4 years ago

Verified 9248e270

[Wav2Vec2FeatureExtractor] smal fixes (#10455)

patil-suraj committed 4 years ago

Verified a106bde5

remove feature extraction config (#10457)

patrickvonplaten committed 4 years ago

Verified 11655faf

Add Fine-Tuning for Wav2Vec2 (#10145)

patrickvonplaten committed 4 years ago

Verified 0234de84

Update ibert.rst (#10445)

patrickvonplaten committed 4 years ago

Verified 3c733f32

Adds terms to Glossary (#10443)

darigovresearch committed 4 years ago

Verified aeba4f95

Introduce save_strategy training argument (#10286)

tanmay17061 committed 4 years ago

Verified 256482ac

updated logging and saving metrics (#10436)

bhadreshpsavani committed 4 years ago

Verified aca6288f

[run_seq2seq.py] restore functionality: saving to test_generations.txt (#10428)

stas00 committed 4 years ago

Verified f52a1589

Fix conda-build (#10431)

LysandreJik committed 4 years ago

Verified 311b7048

[examples] better model example (#10427)

stas00 committed 4 years ago

Verified ee04b698

Ray Tune Integration Bug Fixes (#10406)

amogkam committed 4 years ago

Verified a85eb616

Add Ray Tune hyperparameter search integration test (#10414)

krfricke committed 4 years ago

Verified 98569d4b

[LED] Correct Docs (#10419)

patrickvonplaten committed 4 years ago

Verified d03695f3

Sagemaker Model Parallel tensoboard writing fix (#10403)

mansimane committed 4 years ago

Verified 7fc686ef

[ci, flax] non-existing models are unlikely to pass tests (#10409)

julien-c committed 4 years ago

Verified 83d2d55c

Fix run_glue evaluation when model has a label correspondence (#10401)

sgugger committed 4 years ago

Verified 17b6e0d4

Make Barthez tokenizer tests a bit faster (#10399)

sgugger committed 4 years ago

Verified 26f8b2cb

Fix None in add_token_positions - issue #10210 (#10374)

andreabac3 committed 4 years ago

Verified b040e6ef

Add support for ZeRO-2/3 and ZeRO-offload in fairscale (#10354)

sgugger committed 4 years ago

Verified 9d14be5c

Ignore unexpected weights from PT conversion (#10397)

LysandreJik committed 4 years ago

Verified 88cc26dc

I-BERT model support (#10153)

kssteven418 committed 4 years ago

Verified 63645b3b

[PretrainedFeatureExtractor] + Wav2Vec2FeatureExtractor, Wav2Vec2Processor, Wav2Vec2Tokenizer (#10324)

patrickvonplaten committed 4 years ago

Verified cb38ffcc

Remove unused variable in example for Q&A (#10392)

abhishekkrthakur committed 4 years ago

Verified 9dc78257

Bugfix: Removal of padding_idx in BartLearnedPositionalEmbedding (#10200)

mingruimingrui committed 4 years ago

Verified 894db670

Older