huggingface/transformers

Pull Requests Commits

torch-light: add executorch (+onnx) for CPU exporter tests

tarekziade committed 1 day ago

1b19d6f4

[Fix] Remove deprecated argument from `kernels` call (#47100)

remi-or committed 1 day ago

Verified 3fa92a08

Fix device_map computation when the no_split_modules have different sizes (#47203)

Cyrilvallez committed 1 day ago

Verified 6d960ca0

Remove executorch from all-latest-gpu image + add torch smoke test (#47196)

tarekziade committed 1 day ago

Verified be5e160c

Fix crash in greedy assisted generation with different tokenizers (#46936)

Sunt-ing committed 1 day ago

Verified 8ca0fbec

[Fix] Make DeepGEMM triton fallback more robust (#47126)

remi-or committed 1 day ago

Verified 17a09eeb

[Generation] Add static ensemble verification for lossy speculative decoding (#45979)

kasakh committed 1 day ago

Verified 753de4f6

Add native FSDP2 module + migration (#46707)

3outeille committed 1 day ago

Verified be7a11a8

Fix experts implementation in two spots (#47097)

remi-or committed 1 day ago

Verified ba6f4983

[Fix] Remove old automatic cross attn pattern from output recorders (#47117)

remi-or committed 1 day ago

Verified 4d828999

🌐 [i18n-KO] Translate accelerator_selection.md to Korean (#47157)

kkwjk2718 committed 1 day ago

Verified d45139cf

[i18n-KO] Translate optimum.md to Korean and fix Furiosa typo (#47156)

kkwjk2718 committed 1 day ago

Verified eb36335d

[docs] fix curly quotes rendering to straight quotes (#47135)

clijo committed 2 days ago

Verified 1f2fd058

Fix custom code which doesn't know about the new linear layer type names (#47174)

hmellor committed 2 days ago

Verified 4db1cacb

Reject path traversal in the `transformers_weights` config field (#46890)

LinZiyuu committed 2 days ago

Verified bce8fd08

[docs] Custom code conversion mapping (#47114)

stevhliu committed 2 days ago

Verified 3bbac3e0

Add exporters min version requirements and test skip (#47161)

IlyasMoutawwakil committed 2 days ago

Verified d15c0361

tests: reduce processor test memory usage and use tiny test assets (#47168)

ydshieh committed 2 days ago

Verified 41fc5b47

Clarify input device placement in the Quicktour inference example (#47136)

samyuktahegde committed 2 days ago

Verified b7653886

Extend continuous batching memory prediction test to XPU (#47159)

sywangyi committed 2 days ago

Verified e8cfae1d

Fix case where `_LazyAutoMapping.register` is passed a `str` key (#47148)

hmellor committed 2 days ago

Verified 73cde7c0

[docs] MoE decode switching (#47149)

stevhliu committed 2 days ago

Verified 772f5400

add XPU output expectations for minicpm3 tests (#47092)

kaixuanliu committed 2 days ago

Verified 6f5f5b4d

Diffusion gemma: fix failed test cases (#47025)

kaixuanliu committed 2 days ago

Verified 7a69c513

add XPU Expectation for cosmos3_omni tests (#46880)

kaixuanliu committed 2 days ago

Verified 8e502a17

Fix IndexError Bug in XLMRoberta/Camembert ForMultipleChoice by restoring the pooler (#47147)

pariidanDKE committed 2 days ago

Verified 187fda1b

Restrict QuantizedCache to only full attention (#47160)

Cyrilvallez committed 2 days ago

Verified d876a49f

Skip caching_allocator_warmup on Neuron (no reuse pool to warm; currently OOMs) (#47029)

dacorvo committed 2 days ago

Verified d49affb3

[docs] continuous batching (offloading behavior, max batch tokens, block size minimum) (#46925)

stevhliu committed 3 days ago

Verified 4af901d5

Fix MoE decode optimization for experts living in a submodel (#47107)

IlyasMoutawwakil committed 3 days ago

Verified 1d828044

Older