NielsRogge/transformers

Pull Requests Commits

Remove task guides auto-update in favor of links towards task pages (#30429)

LysandreJik committed 1 year ago

Verified 0eb8fbcd

[`LlamaTokenizerFast`] Refactor default llama (#28881)

ArthurZucker committed 1 year ago

Verified e34da3ee

Fix use_cache for xla fsdp (#30353)

alanwaketan committed 1 year ago

Verified 12c39e56

Rename torch.run to torchrun (#30405)

steven-safeai committed 1 year ago

Verified b8b1e442

Remove old TF port docs (#30426)

Rocketknight1 committed 1 year ago

Verified 696ededd

Fix LayoutLMv2 init issue and doctest (#30278)

ydshieh committed 1 year ago

Verified 416fdbad

FIX: re-add bnb on docker image (#30427)

younesbelkada committed 1 year ago

Verified d179b9dc

Make EosTokenCriteria compatible with mps (#30376)

pcuenca committed 1 year ago

Verified 4b63d013

fix for itemsize => element_size() for torch backwards compat (#30133)

winglian committed 1 year ago

Verified 57fc00f3

Fix on "cache position" for assisted generation (#30068)

zucchini-nlp committed 1 year ago

Verified 77b59dce

Jax: scipy version pin (#30402)

gante committed 1 year ago

Verified 31921d8d

[tests] add `require_torch_sdpa` for test that needs sdpa support (#30408)

faaany committed 1 year ago

Verified 2d61823f

fix: link to HF repo/tree/revision when a file is missing (#30406)

mapmeld committed 1 year ago

Verified 04ac3245

remove redundant logging from longformer (#30365)

riklopfer committed 1 year ago

Verified 179ab098

[Grounding DINO] Add support for cross-attention in GroundingDinoMultiHeadAttention (#30364)

EduardoPach committed 1 year ago

Verified c651ea98

Add inputs embeds in generation (#30269)

zucchini-nlp committed 1 year ago

Verified 408453b4

show `-rs` to show skip reasons (#30318)

ArthurZucker committed 1 year ago

Verified 6c1295a0

[docs] LLM inference (#29791)

stevhliu committed 1 year ago

Verified e74d793a

[FEAT]: EETQ quantizer support (#30262)

dtlzhuangz committed 1 year ago

Verified b4c18a83

Add sdpa and fa2 the Wav2vec2 family. (#30121)

kamilakesbi committed 1 year ago

Verified 569743f5

FIX / PEFT: Pass device correctly to peft (#30397)

younesbelkada committed 1 year ago

Verified 367a0dbd

Fix DETA save_pretrained (#30326)

qubvel committed 1 year ago

Verified 13b3b90a

Jamba: fix left-padding test (#30389)

gante committed 1 year ago

Verified 6c7335e0

Fix layerwise GaLore optimizer hard to converge with warmup scheduler (#30372)

hiyouga committed 1 year ago

Verified f3b3533e

Terminator strings for generate() (#28932)

Rocketknight1 committed 1 year ago

Verified 0d84901c

Update docstrings for text generation pipeline (#30343)

Rocketknight1 committed 1 year ago

Verified 0e9d44d7

`Llama` family, fix `use_cache=False` generation (#30380)

ArthurZucker committed 1 year ago

Verified 2d92db84

Add FSDP config for CPU RAM efficient loading through accelerate (#30002)

helloworld1 committed 1 year ago

Verified f16caf44

GenerationConfig: warn if pad token is negative (#30187)

zucchini-nlp committed 1 year ago

Verified 91389357

Enable multi-device for more models (#30379)

jla524 committed 1 year ago

Verified 8b02bb6e

Newer Older