younesbelkada/transformers

Pull Requests Commits

younesbelkada committed 3 years ago

fcbcfd34

Merge remote-tracking branch 'upstream/main' into blip-text-refactor-test

younesbelkada committed 3 years ago

37386eb4

younesbelkada committed 3 years ago

3e8314dc

remove test as duplicate

younesbelkada committed 3 years ago

44e66445

fix FSDP version related issues (#22489)

pacman100 committed 3 years ago

Verified ee8e80a0

Update tiny model summary file for recent models (#22637)

ydshieh committed 3 years ago

Verified c7ec71ba

[`Blip`] Fix slow tests and doctests with correct values (#22632)

younesbelkada committed 3 years ago

Verified ed672864

LlamaTokenizerFast Fix (.., from_slow=True). (#22630)

Narsil committed 3 years ago

Verified 6a02e980

[`bnb`] 8bit models should not be converted to `DDP` (#22628)

younesbelkada committed 3 years ago

Verified 09a9888f

A script to add/update `pipeline_model_mapping` systematically (#22180)

ydshieh committed 3 years ago

Verified d0b83fe2

update_pip_test_mapping (#22606)

ydshieh committed 3 years ago

Verified fa01127a

docs: Fix broken link to generation strategies (#22623)

Connor Henderson committed 3 years ago

Verified 321b0908

Make tiny model creation + pipeline testing more robust (#22500)

ydshieh committed 3 years ago

Verified 2c22bc79

Backbone add mixin tests (#22542)

amyeroberts committed 3 years ago

Verified 12d51db2

Seq2SeqTrainer: use unwrapped model to retrieve the generation config (#22584)

gante committed 3 years ago

Verified 48706c71

Revert error back into warning for byte fallback conversion. (#22607)

Narsil committed 3 years ago

Verified 0aa1153f

Adding Llama FastTokenizer support. (#22264)

Narsil committed 3 years ago

Verified 1670be4b

feat(model parallelism): moving the labels to the same device as the logits for gpt2 and bart (#22591)

kausmeows committed 3 years ago

Verified 15641892

Use native TF checkpoints for the BLIP TF tests (#22593)

Rocketknight1 committed 3 years ago

Verified e577bd0f

Add DePlot + MatCha on `transformers` (#22528)

younesbelkada committed 3 years ago

Verified 176ceff9

Adding support for BPE merge creation from scores instead of ids. (#22582)

Narsil committed 3 years ago

Verified 126eafe3

Fix a typo in one of the BLIP pretrained checkpoint names (#22588)

Rocketknight1 committed 3 years ago

Verified 12f1a3bb

Sync preprocesses before loading the processor at run_speech_recognition_ctc.py (#21926)

mpenagar committed 3 years ago

Verified d5239bab

docs: ko: complete `_toctree.yml` (#22581)

wonhyeongseo committed 3 years ago

Verified f49b0762

Add thousands separator in training summary (#22583)

qmeeus committed 3 years ago

Verified 4861c258

Fix PT-TF equivalence test for GPT1 (#22586)

Rocketknight1 committed 3 years ago

Verified 2a91a9ef

Tests: disable `accelerate_tests` mark warnings (#22585)

gante committed 3 years ago

Verified 06842849

Move back doctest instructions to setup.cfg (#22587)

sgugger committed 3 years ago

Verified 6c640f09

Generate: `TextIteratorStreamer` timeout (#22576)

gante committed 3 years ago

Verified 861ff890

Skip failing test

sgugger committed 3 years ago

Verified 11fd2c77

Older