huggingface/transformers

Pull Requests Commits

ydshieh committed 1 year ago

d4044835

Refactor the way we handle outputs for new llamas and new models (#39120)

ArthurZucker committed 1 year ago

Verified ca7e1a37

Update expected values (after switching to A10) - part 8 - Final (#39220)

ydshieh committed 1 year ago

Verified e6a8063e

Update expected values (after switching to A10) - part 7 (#39218)

ydshieh committed 1 year ago

Verified cd8a041a

Add packed tensor format support for flex/sdpa/eager through the mask! (#39194)

Cyrilvallez committed 1 year ago

Verified 0cf27916

Update expected values (after switching to A10) - part 6 (#39207)

ydshieh committed 1 year ago

Verified 037755ed

Update expected values (after switching to A10) - part 5 (#39205)

ydshieh committed 1 year ago

Verified 1168f57a

Fix continuous batching in `transformers serve` (#39149)

LysandreJik committed 1 year ago

Verified 7d9e52f3

[serve] Cursor support, move docs into separate page, add more examples (#39133)

gante committed 1 year ago

Verified 85d93cc6

[typing] better return typehints for `from_pretrained` (#39184)

qubvel committed 1 year ago

Verified e15b06d8

Update expected values (after switching to A10) - part 4 (#39189)

ydshieh committed 1 year ago

Verified a25fc359

[`Dia`] Change ckpt path in docs (#39181)

vasqu committed 1 year ago

Verified b31e9d19

Fix many HPU failures in the CI (#39066)

IlyasMoutawwakil committed 1 year ago

Verified 18e0cae2

Decouple device_map='auto' and tp_plan='auto' (#38942)

SunMarc committed 1 year ago

Verified bff964c4

when delaying optimizer creation only prepare the model (#39152)

winglian committed 1 year ago

Verified 8178c431

[glm4v] fix video inference (#39174)

zucchini-nlp committed 1 year ago

Verified 91221da2

Test fixes for Aria (and some Expectation for llava_next_video) (#39131)

remi-or committed 1 year ago

Verified ebfbcd42

Update expected values (after switching to A10) - part 3 (#39179)

ydshieh committed 1 year ago

Verified 37a239ca

Update expected values (after switching to A10) - part 2 (#39165)

ydshieh committed 1 year ago

Verified 9326fc33

Random serve fixes (#39176)

pcuenca committed 1 year ago

Verified 25cd65ac

[serve] Model name or path should be required (#39178)

LysandreJik committed 1 year ago

Verified 548794b8

[generate] document non-canonical beam search default behavior (#39000)

gante committed 1 year ago

Verified 2d561713

[docs] ViTPose (#38630)

stevhliu committed 1 year ago

Verified df12d87d

Reduce Glm4v model test size significantly (#39173)

Cyrilvallez committed 1 year ago

Verified 2b4a12b5

Fix missing initializations for models created in 2024 (#38987)

bvantuan committed 1 year ago

Verified e355c0a1

Blip2 fixes (#39080)

remi-or committed 1 year ago

Verified 1125513a

Fix multimodal processor get duplicate arguments when receive kwargs for initialization (#39125)

Isotr0py committed 1 year ago

Verified 28df7f85

🚨🚨🚨 [eomt] make EoMT compatible with pipeline (#39122)

yaswanth19 committed 1 year ago

Verified b61023a1

[smolvlm] fix video inference (#39147)

zucchini-nlp committed 1 year ago

Verified 4d5822e6

fix default value of config to match checkpionts in LLaVa-OV models (#39163)

ved1beta committed 1 year ago

Verified 9b2f5b66

Older