huggingface/transformers

Pull Requests Commits

`td`: follow the daemon's real log, and make bare `td` tail it

ArthurZucker committed 36 days ago

5db22cde

`td`: rank 0 to the console, job durations, and a --detach mode

ArthurZucker committed 36 days ago

d19e9e9c

Make `td` recover instead of hanging, and free GPU memory between jobs

ArthurZucker committed 36 days ago

461d017a

Add `td`, a persistent multi-node job runner

arthur@huggingface.co committed 36 days ago

51654906

Kernel api doc, part 2 (#46889)

michaelbenayoun committed 37 days ago

Verified 7219d282

[InternVL] Normalize num_patches before np.cumsum (#48469)

lorenzozanee committed 37 days ago

Verified 9ab9957d

Fix Glm4vMoeIntegrationTest: offload_folder + MemoryCleanupMixin (#48776)

ydshieh committed 37 days ago

Verified f24b457b

Fix silently random-initializing `RTDetrModel`/`SEWDForCTC` loads (wrong `base_model_prefix`) (#48744)

abhinav-phi committed 37 days ago

Verified 56c5e876

fix videomae load error (#48675)

sywangyi committed 37 days ago

Verified 06a7a68e

Switch daily CI to torch 2.14 — update expected outputs (#48750)

ydshieh committed 37 days ago

Verified a2c15b30

Enable FSDP2 + expert parallelism via a 2-D (fsdp, tp) device mesh (#48516)

qgallouedec committed 38 days ago

Verified 415e6d2f

unifying device_mesh init to enable PP + TP inference (#48155)

3outeille committed 38 days ago

Verified 5474a55e

Use huggingface_hub httpx export (#48685)

Wauplin committed 39 days ago

Verified 80a79d1b

[DeepseekV3] OOM cascade root-cause investigation (generator ref leak in conversion_mapping) (#48720)

ydshieh committed 39 days ago

Verified df04b012

Keep the expert-parallel sentinel slots out of the router gradient (#48689)

qgallouedec committed 39 days ago

Verified 67f50855

add HyperClovaX Vision (#44314)

jp1924 committed 39 days ago

Verified a6cf156d

Add flash_attention_4 in attn_implementation AutoModel docstring (#48684)

3manifold committed 39 days ago

Verified 9c944e97

[CI] Replace hardcoded username allowlists with author_association check in workflow triggers (#48712)

ydshieh committed 39 days ago

Verified 3f601734

Fix D-FINE / RT-DETR main loss being computed over the denoising queries (#48528)

stefan-it committed 39 days ago

Verified 846bdfdd

[serge] Fix 1 integration test for model `flex_olmo` failing with `other` (other (1)) (#48668)

sergereview[bot] committed 40 days ago

Verified 6b07e451

QA: Added a `MemoryCleanupMixin` class for tests (#48681)

tarekziade committed 40 days ago

Verified 8fcb5239

Fix slow integration tests on XPU (#48611)

jiqing-feng committed 40 days ago

Verified 5509c70d

Fix Qwen2.5-VL temporal RoPE for fractional video intervals (#48669)

yeyeyeping committed 40 days ago

Verified bd15bc95

[serge] Fix 2 integration tests for model `mistral` failing with `other` (other (2)) (#48429)

sergereview[bot] committed 40 days ago

Verified 177e90dd

[serge] Fix 2 integration tests for model `minimax` failing with `output_mismatch` (tensor values differ (2)) (#48515)

sergereview[bot] committed 40 days ago

Verified ff4bb5ee

[Fix] yolos offload issue (#48688)

molbap committed 40 days ago

Verified 9d13f8cf

Skip the expert-parallel sentinel masking when expert parallelism is off (#48201)

qgallouedec committed 40 days ago

Verified 606e6e8b

Synthetic test assets (#48589)

tarekziade committed 40 days ago

Verified c6947074

Fix deepspeed ci (#48640)

SunMarc committed 40 days ago

Verified 29723434

[MusicgenMelody] Fix conditioning silently dropped at generation step 0 (#48679)

ydshieh committed 40 days ago

Verified 5bed6296

Older