Commits huggingface/transformers

Let the GPU verify caller turn on the memory probe (#48001)

tarekziade committed 32 days ago

Verified ec30f2f8

Fix MTP config when mlp_layer_types is absent (#48015)

Cyrilvallez committed 32 days ago

Verified 298ae854

Remove duplicate block_sparse_moe assignment in GraniteMoeDecoderLayer (#47876)

Aman2394 committed 32 days ago

Verified 28317985

Bump default flash-attn2 hub kernel version to v3 (#47863)

jiqing-feng committed 32 days ago

Verified 6724a56b

Align logit distributions for CandidateGenerators using sampling (#48007)

Cyrilvallez committed 32 days ago

Verified 240b8330

[ModernVBERT] Fix integration test checkpoint (404 since April) (#48009)

ydshieh committed 32 days ago

Verified e378772d

Fix cropping (#48006)

Cyrilvallez committed 32 days ago

Verified a93de5e1

Fix DFlash candidate token device mismatch with device_map="auto" (#47877)

sywangyi committed 32 days ago

Verified 18388f3b

:red_circle: Allow tokenizers 0.23.1 (#46381)

ArthurZucker committed 33 days ago

Verified 242f5df9

[Whisper] Fix speculative decoding: UnboundLocalError, cache corruption, and speed regression (#48000)

ydshieh committed 33 days ago

Verified a5f41e09

[new model] step 3.7 (#46658)

itazap committed 33 days ago

Verified c49429ed

[Whisper] Fix batch decode_with_timestamps in WhisperTokenizer.decode() (#47997)

ydshieh committed 33 days ago

Verified 64b6595f

[Whisper] Fix integration test failures on A10G (dtype, stale values, API changes) (#47995)

ydshieh committed 33 days ago

Verified 16cbed7e

[DeepSeekV2] Fix integration tests OOM: use device_map=auto instead of 8-bit quantization (#47991)

ydshieh committed 34 days ago

Verified a61d5f9e

[OLMoE] Update expected logits for A10G and add torch.no_grad() (#47989)

ydshieh committed 34 days ago

Verified bd5df99a

[OLMo] Fix OOM in logits tests by adding torch.no_grad() (#47986)

ydshieh committed 34 days ago

Verified c7ad7287

[GPTNeoX] Fix post_processor not overridden when loading from pretrained (OLMo garbage generation) (#47988)

ydshieh committed 34 days ago

Verified ce5c8f5e

[AXK1] Fix expected logits for CUDA A10G (#47980)

ydshieh committed 35 days ago

Verified 96fe6dce

[Gemma] Update expected values for A10G (#47976)

ydshieh committed 35 days ago

Verified df0db63c

[docs] torchcodec + vision processor features (#47935)

stevhliu committed 35 days ago

Verified a597f974

Fix gemma4 video to device (#47896)

guarin committed 35 days ago

Verified c1ff1186

[emu3] 🦮 Black Labrador is back! Fix image generation broken since #37033 (#47948)

ydshieh committed 35 days ago

Verified 60eefbc8

Remove stale (None, None) fallback in qwen2_5_vl batch_different_resolutions test (#47972)

ydshieh committed 35 days ago

Verified 7acb594b

Fix CLIP _init_weights when a child module carries quantized weights (#47921)

Bluear7878 committed 35 days ago

Verified 13190137

Potential fix for code scanning alert no. 267: Artifact poisoning (#47949)

tarekziade committed 35 days ago

Verified 463f684a

Fix GatedDeltaNet A_log dtype to prevent -inf under bfloat16 init (#47944)

Nkluge-correa committed 35 days ago

Verified 95940bf8

[serge] Fix 2 integration tests for model `got_ocr2` failing with `other` (other (2)) (#47937)

sergereview[bot] committed 35 days ago

Verified ed3edd87

Fix Jinja block endings in CHAT WITH MODELS' Writing a chat template … (#47960)

ak1for2business-prog committed 35 days ago

Verified 6d2a4212

[tests] Fix expected output for Qwen2.5-VL batch_wo_image on CUDA (#47968)

ydshieh committed 35 days ago

Verified 53a4389e

Declare sdpa support in `TimmWrapper` (#47939)

jiqing-feng committed 35 days ago

Verified 1cdaf598