huggingface/transformers

Pull Requests Commits

use a deterministic approach

tarekziade committed 9 hours ago

3cb2c2ea

tarekziade committed 9 hours ago

c9ad6604

different approach to make sure loss exists

tarekziade committed 9 hours ago

a6c3a616

:rotating_light: Modify ModernBERT's default attention implementation to stop using FA (#43764)

YangKai0616 committed 12 hours ago

Verified 2db2c887

Fix-release-ubild (#43773)

ArthurZucker committed 15 hours ago

Verified 0b2900dd

Add VibeVoice Acoustic Tokenizer (#43400)

ebezzam committed 15 hours ago

Verified 281eeef1

Improved `AGENTS.md` (#43763)

tarekziade committed 15 hours ago

Verified ab3d1c22

Bump dev version (#43777)

qgallouedec committed 16 hours ago

Verified ecd0536d

Fix `convert_rope_params_to_dict` so it uses `rope_theta` from the config (#43766)

hmellor committed 16 hours ago

Verified 438856e4

[Trainer] Move NEFTune impl to standalone functions (#43714)

SunMarc committed 1 day ago

Verified 393b4b3d

[docs] deploying (#43241)

stevhliu committed 1 day ago

Verified b92f8ffc

[Trainer] Move sort and rotate checkpoints to standalone functions (#43736)

SunMarc committed 1 day ago

Verified 759b5c8f

Fix EP post merge (#43730)

ArthurZucker committed 1 day ago

Verified e1f37667

stablize the expect output for eomt_dinov3 model test (#43758)

kaixuanliu committed 1 day ago

Verified 35b85bd5

fix(models): Migrate legacy segmentation_indices to out_indices in BeitConfig (#43505)

harshaljanjani committed 1 day ago

Verified b98123c4

[docs] Update torch version (#42135)

stevhliu committed 1 day ago

Verified ffaf03b3

Remove SDPA workarounds for torch 2.4+ (#43754)

cyyever committed 1 day ago

Verified cf763251

add use_deterministic to guarantee the consistency for youtu-llm model (#43759)

kaixuanliu committed 1 day ago

Verified 239c3d6c

fix: add compatible_model_types to suppress model type mismatch warnings (#43495)

leoneperdigao committed 1 day ago

Verified a8e791db

Fix T5 v1.1 detection (#43681)

githubnemo committed 1 day ago

Verified b8a1c69b

Add moonshine streaming (#43702)

eustlb committed 2 days ago

Verified ace7c370

Allow bi-directional attention for all models (#43705)

Cyrilvallez committed 2 days ago

Verified 83bce8d8

Docs: fix Training step by removing tokenizer from trainer initialization (#43733)

nesjett committed 2 days ago

Verified 452c179e

Fix scheduler initialization order (#43711)

SunMarc committed 2 days ago

Verified 1744f8f3

Fix accelerate integration import (#43732)

SunMarc committed 2 days ago

Verified 8dce3100

🚨 T5Gemma2 model structure (#43633)

zucchini-nlp committed 2 days ago

Verified d75266f1

Update torch minimum version to 2.4 (#41307)

cyyever committed 2 days ago

Verified a276a763

Fix dtype in image-text-to-text pipe (#43731)

zucchini-nlp committed 2 days ago

Verified d95672f8

Preventing initialization of siglip's lecun_normal_, default_flax_embed_init in ZeRO3 (#43574)

jp1924 committed 2 days ago

Verified 225254c1

🚨 Generation cache preparation (#43679)

zucchini-nlp committed 2 days ago

Verified 1687954f

Older