huggingface/transformers

Pull Requests Commits

add other models

itazap committed 1 year ago

38a8ce45

itazap committed 1 year ago

0a246a5a

itazap committed 1 year ago

0d72f204

rm func from some models

itazap committed 1 year ago

1cb4b7fb

rm build_input function for fast, don't test prepare_for_model for fast as we don't use it

itazap committed 1 year ago

f3fb6164

rm build_input.. from old file

itazap committed 1 year ago

2b8774a7

rm build_inputs_with_special_tokens from llama and gemma

itazap committed 1 year ago

adeb8cdd

itazap committed 1 year ago

148e3159

itazap committed 1 year ago

cc76a4f1

itazap committed 1 year ago

0202f862

[MODEL] Add Falcon H1 (#38249)

younesbelkada committed 1 year ago

Verified 6829936e

tp plan should not be NONE (#38255)

ArthurZucker committed 1 year ago

Verified e288ee00

Revert parallelism temporarily (#38240)

LysandreJik committed 1 year ago

Verified 711d78d1

CI reporting improvements (#38230)

ydshieh committed 1 year ago

Verified feec294d

Protect ParallelInterface

LysandreJik committed 1 year ago

cb513e35

LysandreJik committed 1 year ago

f4ef41c4

[gemma3] fix bidirectional attention mask (#38080)

zucchini-nlp committed 1 year ago

Verified f834d368

[mllama] fix loading and inference (#38223)

zucchini-nlp committed 1 year ago

Verified 2edb0e4b

Add padding-free to bamba (#35861)

garrett361 committed 1 year ago

Verified 390f1534

Fixing Bitnet after use_rms_norm introduction (#38229)

MekkCyber committed 1 year ago

Verified 2a794713

Enable Quantize KV Cache for Mistral Model (#35042)

Bojun-Feng committed 1 year ago

Verified 96618960

parallelism goes brrr (#37877)

NouamaneTazi committed 1 year ago

Verified 1c2f36b4

Fix Llama4 (#38222)

Cyrilvallez committed 1 year ago

Verified b591d925

Mamba2 remove unecessary test parameterization (#38227)

ivarflakstad committed 1 year ago

Verified 3f0b7d0f

Minor llama4 fixes (#38123)

molbap committed 1 year ago

Verified 9cde2f5d

fix dead flax links modeling_flax_pytorch_utils.py (#38212)

dizer-ti committed 1 year ago

Verified 856f034f

Make `train_dataset` attribute in `_get_train_sampler` optional (#38226)

SunMarc committed 1 year ago

Verified bb3c6426

In Llama4 fix wrongly inverted causal attention mask when using SDPA implementation (#38094)

sogartar committed 1 year ago

Verified 2ad152f8

Disable torchscript tests for AriaForConditionalGenerationModelTest (#38225)

ivarflakstad committed 1 year ago

Verified de70c842

Add support to Marimo Notebooks and Enverge.ai (#38210)

brenoca committed 1 year ago

Verified 8ea61c45

Older