huggingface/transformers

Pull Requests Commits

Merge commit 'ed2f65a9' into dynamic_length_on_75bbfd5b

ydshieh committed 2 years ago

c302fff6

Merge commit 'cbf98115' into dynamic_length_on_75bbfd5b

ydshieh committed 2 years ago

e4859e6e

Merge commit '10bc8b4e' into dynamic_length_on_0ae789e0

ydshieh committed 2 years ago

ed2f65a9

0ae789e0 + dynamic length in static cache

ydshieh committed 2 years ago

cbf98115

fix EosTokenCriteria

ydshieh committed 2 years ago

10bc8b4e

Cache: Static cache as a standalone object (#30476)

gante committed 2 years ago

Verified 75bbfd5b

Enable multi-device for more models (#30409)

jla524 committed 2 years ago

Verified 0ae789e0

Pass `use_cache` in kwargs for GPTNeoX (#30538)

zucchini-nlp committed 2 years ago

Verified c712d05a

Include safetensors as part of `_load_best_model` (#30553)

muellerzr committed 2 years ago

Verified a3aabc70

Reenable SDPA's FA2 During Training with torch.compile (#30442)

warner-benjamin committed 2 years ago

Verified 9df8b301

Fix repo. fetch/checkout in PR slow CI job (#30537)

ydshieh committed 2 years ago

Verified 87be06ca

Update runner tag for PR slow CI (#30535)

ydshieh committed 2 years ago

Verified c0242188

Fix broken link to Transformers notebooks (#30512)

clinty committed 2 years ago

Verified bdbe1662

Pass attn_implementation when using AutoXXX.from_config (#30507)

amyeroberts committed 2 years ago

Verified e8acb700

Allow boolean FSDP options in fsdp_config (#30439)

helloworld1 committed 2 years ago

Verified 80126f98

Fix link in dbrx.md (#30509)

eitanturok committed 2 years ago

Verified 73014b56

[SegGPT] Fix seggpt image processor (#29550)

EduardoPach committed 2 years ago

Verified 6d4cabda

load_image - decode b64encode and encodebytes strings (#30192)

amyeroberts committed 2 years ago

Verified c793b26f

Fix GroundingDINO, DPR after BERT SDPA update (#30506)

amyeroberts committed 2 years ago

Verified e7d52a10

[examples] update whisper fine-tuning (#29938)

sanchit-gandhi committed 2 years ago

Verified 38b53da3

[`DETR`] Remove timm hardcoded logic in modeling files (#29038)

amyeroberts committed 2 years ago

Verified aafa7ce7

Remove skipping logic now that set_epoch exists (#30501)

muellerzr committed 2 years ago

Verified 77ff304d

[`BERT`] Add support for sdpa (#28802)

hackyon committed 2 years ago

Verified dfa7b580

Use the Keras set_random_seed in tests (#30504)

Rocketknight1 committed 2 years ago

Verified 2de5cb12

Update `dtype_byte_size` to handle torch.float8_e4m3fn/float8_e5m2 types (#30488)

mgoin committed 2 years ago

Verified 20081c74

Fix the `bitsandbytes` error formatting ("Some modules are dispatched on ...") (#30494)

kyo-takano committed 2 years ago

Verified 59e715f7

FEAT: PEFT support for EETQ (#30449)

younesbelkada committed 2 years ago

Verified 19cfdf0f

[docs] Spanish translation of pipeline_tutorial.md (#30252)

aaronjimv committed 2 years ago

Verified a98c4179

Quantization: `HfQuantizer` quant method update (#30484)

younesbelkada committed 2 years ago

Verified 26ddc580

Add sidebar tutorial for chat models (#30401)

Rocketknight1 committed 2 years ago

Verified f3962712

Older