huggingface/transformers

Pull Requests Commits

Optimize t5 tokenize logic to avoid redundant calls (#32270)

leejet committed 1 year ago

Verified 5019aabf

Upload new model failure report to Hub (#32264)

ydshieh committed 1 year ago

Verified f2122cc6

🚨 Bloom support for cache class (#31445)

zucchini-nlp committed 1 year ago

Verified f7396876

Llama 3.1: replace for loop by tensor ops at inv_freq initialization (#32244)

gante committed 1 year ago

Verified 44f6fdd7

More flexible trigger condition (#32251)

ydshieh committed 1 year ago

Verified 8da90687

Flash-Attn: fix generation when no attention mask or no pading (#32241)

zucchini-nlp committed 1 year ago

Verified 81233c06

[tests] fix `static` cache implementation is not compatible with `attn_implementation==flash_attention_2` (#32039)

faaany committed 1 year ago

Verified 27c7f971

Add check for `target_sizes is None` in `post_process_image_guided_detection` for owlv2 (#31934)

catalys1 committed 1 year ago

Verified 5f841c74

Adds: extra_repr for RMSNorm layers in most models (#32204)

rohitdwivedula committed 1 year ago

Verified f9756d9e

Refactor: Removed un-necessary `object` base class (#32230)

Sai-Suraj-27 committed 1 year ago

Verified b8e5cd53

don't log base model architecture in wandb if log model is false (#32143)

joaonadkarni committed 1 year ago

Verified 1c7ebf1d

Resize embeds with DeepSpeed (#32214)

zucchini-nlp committed 1 year ago

Verified c46edfb8

Llava: generate without images (#32183)

zucchini-nlp committed 1 year ago

Verified fad15fba

Generation: stop at `eos` for assisted decoding (#31301)

zucchini-nlp committed 1 year ago

Verified 4ab33c2d

Fix code snippet for Grounding DINO (#32229)

qubvel committed 1 year ago

Verified 9d6c0641

Allow a specific microphone to be used by the ffmpeg audio pipeline utility functions. Default to using the currently active microphone on Mac (#31846)

jrhe committed 1 year ago

Verified 3a83ec48

translate philosophy.md to chinese (#32177)

ji-huazhong committed 1 year ago

Verified 6ed0bf1e

Follow up for #31973 (#32025)

ydshieh committed 1 year ago

Verified df6eee92

[warnings] fix E721 warnings (#32223)

kashif committed 1 year ago

Verified de231889

[BigBird Pegasus] set _supports_param_buffer_assignment to False (#32222)

kashif committed 1 year ago

Verified 9b9a54e6

Update question_answering.py (#32208)

avlewis committed 1 year ago

Verified 1ecedf1d

remove unnecessary guard code related with pytorch versions 1.4.2 ~ 1.7.0 (#32210)

ji-huazhong committed 1 year ago

Verified f53a5dec

[whisper] fix short-form output type (#32178)

sanchit-gandhi committed 1 year ago

Verified 5658e749

fix: Replaced deprecated `unittest method` with the correct one (#32198)

Sai-Suraj-27 committed 1 year ago

Verified 85a1269e

:rotating_light: No more default chat templates (#31733)

Rocketknight1 committed 1 year ago

Verified edd68f4e

Support dequantizing GGUF FP16 format (#31783)

PenutChen committed 1 year ago

Verified 1c122a46

Fix float8_e4m3fn in modeling_utils (#32193)

SunMarc committed 1 year ago

Verified af0e4b7b

Fix resize embedding with Deepspeed (#32192)

zucchini-nlp committed 1 year ago

Verified 1392a686

let's not warn when someone is running a forward (#32176)

ArthurZucker committed 1 year ago

Verified 8d2534c4

RoPE: relaxed rope validation (#32182)

gante committed 1 year ago

Verified e0182f3b

Newer Older