younesbelkada/transformers

Pull Requests Commits

Apply suggestions from code review

younesbelkada committed 3 years ago

Verified d293c254

Update src/transformers/modeling_utils.py

younesbelkada committed 3 years ago

Verified 48b7f8f4

younesbelkada committed 3 years ago

fc1411eb

younesbelkada committed 3 years ago

a59b6381

younesbelkada committed 3 years ago

f03d36e8

address last comments

younesbelkada committed 3 years ago

9042365e

Update src/transformers/utils/quantization_config.py

younesbelkada committed 3 years ago

Verified eda6d401

remove unused function

younesbelkada committed 3 years ago

9fadcf78

younesbelkada committed 3 years ago

42167957

Merge remote-tracking branch 'upstream/main' into serialize-8bit

younesbelkada committed 3 years ago

379b7f39

adapt from suggestions

younesbelkada committed 3 years ago

897cde9f

Relax `eos_token_id < 0` checks in `generate()` from `ValueError` to warning (#22472)

lewtun committed 3 years ago

Verified da68fd69

(Re-)Enable Nightly + Past CI (#22393)

ydshieh committed 3 years ago

Verified 0fe6c6bd

Docs fix: Multinomial sampling decoding needs "num_beams=1", since by default it is usually not 1. (#22473)

manueldeprada committed 3 years ago

Verified d5de578c

Llama: support for `max_position_embeddings` (#22471)

gante committed 3 years ago

Verified 165dd6dc

[NLLB-MoE] `model_type` update for auto mapping (#22470)

ArthurZucker committed 3 years ago

Verified 349e1242

Guard imports of PreTrainedTokenizerFast on is_tokenizers_available (#22285)

Roy Hvaara committed 3 years ago

Verified 11426641

🚨🚨🚨 Fix ordering of height, width for BLIP image processor (#22466)

amyeroberts committed 3 years ago

Verified 4d7a5b5b

Generate: basic token streaming (#22449)

gante committed 3 years ago

Verified 228792a9

Skip flaky NLLB Moe test for now (#22463)

amyeroberts committed 3 years ago

Verified f0aeb1be

Rescale image back if it was scaled during PIL conversion (#22458)

amyeroberts committed 3 years ago

Verified 154c6bb7

Move common properties to BackboneMixin (#21855)

amyeroberts committed 3 years ago

Verified c15f9375

Update: ignore padding support for TransfoXL training when n_clusters==0 (#22457)

StefanHeng committed 3 years ago

Verified cd73b9a8

Pin ruff (#22455)

sgugger committed 3 years ago

Verified 2194943a

Update release instructions (#22454)

sgugger committed 3 years ago

Verified 4c295a26

Avoid using personal HF token in CI (#22453)

ydshieh committed 3 years ago

Verified 97440e9c

Update Neptune docs (#22452)

Sabine committed 3 years ago

Verified 173193cc

Revert "Fix --bf16 option support for Neuron after PR #22300" (#22451)

jeffhataws committed 3 years ago

Verified 5e89a435

[`Pix2Struct`] Fix slow test (#22448)

younesbelkada committed 3 years ago

Verified b844f8a9

Revert "Error (also in original) model, scaling only q matrix not qk.T dot product (qk.T/sqrt(dim_per_head))" (#22444)

sgugger committed 3 years ago

Verified 55dae94c

Older