huggingface/transformers

Pull Requests Commits

Merge branch 'fix_fbgemm_tp' into merging_to_test

MekkCyber committed 280 days ago

7960f1d9

MekkCyber committed 280 days ago

2024e3b5

Merge branch 'fix_fbgemm_tp' into merging_to_test

MekkCyber committed 280 days ago

091657e2

MekkCyber committed 280 days ago

634c9c66

Merge branch 'flex-fix-regression' into merging_to_test

MekkCyber committed 280 days ago

4500320a

MekkCyber committed 280 days ago

b3e08ec1

Preserve requires_grad in pre quantized model (#37354)

jerryzh168 committed 281 days ago

Verified 08e3217b

Cyrilvallez committed 281 days ago

Verified c6f442ae

:rotating_light: :rotating_light: Setup -> setupclass conversion (#37282)

Rocketknight1 committed 281 days ago

Verified 4d0de5f7

fix(qwen): fix shape error when using tp (#36947)

KimmiShi committed 281 days ago

Verified c15a7adb

prune LM Head for USD (#36695)

jmamou committed 281 days ago

Verified 121f91d3

[core] remove `GenerationMixin` inheritance by default in `PreTrainedModel` (#37173)

gante committed 281 days ago

Verified 4321b064

Skip non-selected experts for mixtral and qwen2_moe (#32429)

Coco58323 committed 281 days ago

Verified aab08783

the fix that did not get in

ArthurZucker committed 281 days ago

c7d1e731

[llama 4] dynamic rope decorator (#37365)

gante committed 281 days ago

Verified 35f0f5b5

Set vision config to None for Gemma 1B conversion (#37366)

RyanMullins committed 281 days ago

Verified 530322cc

fix deepspeed job (#37284)

ydshieh committed 281 days ago

Verified 8064cd9b

A bit of cleaning 🧹🧹 (#37215)

Cyrilvallez committed 281 days ago

Verified cdfb018d

Use Python 3.9 syntax in tests (#37343)

cyyever committed 281 days ago

Verified 1e6b546e

convert float for yarn related arguments in rope_scaling (#37139)

bzantium committed 281 days ago

Verified 0fc683d1

Expose blip2qformer (#37254)

alex-jw-brooks committed 281 days ago

Verified 2515a5a2

Multiple llama4 fixe (#37353)

ArthurZucker committed 281 days ago

Verified 2da82e43

Fixing flex attention for torch=2.6.0 (#37285)

salmanmohammadi committed 282 days ago

Verified 794fde7b

more fixes for post-training llama4 (#37329)

winglian committed 282 days ago

Verified b54c2f46

Remove unnecessary attr assignment (#36837)

tugsbayasgalan committed 282 days ago

Verified 754a370b

Updated Model-card for donut (#37290)

Logeswaran7 committed 282 days ago

Verified 31a62c2e

Add bnb to the list of supported quantization methods for LLama4 (#37348)

MekkCyber committed 282 days ago

Verified f8301051

Update Model Card for Jamba (#37152)

ParagEkbote committed 282 days ago

Verified e2b0224d

Improvements in Gemma2 model card (#37076)

devesh-2002 committed 282 days ago

Verified 6cc109c3

Clean up the compressed-tensors integration (#37349)

MekkCyber committed 282 days ago

Verified 8bbcdf54

Older