huggingface/transformers

Pull Requests Commits

feat(serve): add memory admission param to `add_request`

McPatate committed 101 days ago

Verified 69d0e232

test(serve): add `ConversationCacheManager` unittest class

McPatate committed 101 days ago

Verified f76bd410

fix(serve): only evict entries that are not in use

McPatate committed 101 days ago

Verified 668c8b85

refactor(serve): rename ttl class and add docs

McPatate committed 101 days ago

Verified 25e20d72

refactor: explicit `ttl`

McPatate committed 101 days ago

Verified 699e129e

refactor(serve): use `cancel_request` instead of evict req from cache method

McPatate committed 101 days ago

Verified 2daaefa7

fix: add entry lease per cache entry to avoid evicting when active

McPatate committed 101 days ago

Verified 4d7044ce

feat: kv cache retention across conversations

McPatate committed 101 days ago

Verified e524d6e4

Minor fix for #40727 (#40929)

ydshieh committed 101 days ago

Verified 89949c5d

Adding activation kernels (#40890)

MekkCyber committed 101 days ago

Verified c830fc12

[torchao safetensors] renaming get_state_dict function (#40774)

liangel-02 committed 101 days ago

Verified f6999b00

Fix #40067: Add dedicated UMT5 support to GGUF loader (config, tokenizer, test) (#40218)

akshay-babbar committed 101 days ago

Verified 8428c7b9

[Llama4] Remove `image_sizes` arg and deprecate `vision_feature_layer` (#40832)

yaswanth19 committed 101 days ago

Verified ddd4caf0

Processor load with multi-processing (#40786)

zucchini-nlp committed 101 days ago

Verified b82cd1c2

[Docs] Adding documentation of MXFP4 Quantization (#40885)

ariG23498 committed 101 days ago

Verified 6e50a8af

Fix dtype in Paligemma (#40912)

zucchini-nlp committed 101 days ago

Verified cccef4be

🔴Make `center_crop` fast equivalent to slow (#40856)

yonigozlan committed 101 days ago

Verified beb09cbd

[generate] misc fixes (#40906)

gante committed 101 days ago

Verified d4af0d9f

[gemma3] `Gemma3ForConditionalGeneration` compatible with assisted generation (#40791)

gante committed 101 days ago

Verified 3b3f6cd0

disable `test_fast_is_faster_than_slow` (#40909)

ydshieh committed 101 days ago

Verified 88ba0f10

Remove `runner_map` (#40880)

ydshieh committed 101 days ago

Verified 270da897

Improve module name handling for local custom code (#40809)

XuehaiPan committed 101 days ago

Verified df03fc1f

remove dummy EncodingFast (#40864)

cyyever committed 101 days ago

Verified 96bc19bc

Add Olmo3 model (#40778)

2015aroras committed 102 days ago

Verified d0af4269

Set seed for `Glm4vIntegrationTest` (#40905)

ydshieh committed 102 days ago

Verified 65f9ede3

[cache] Only use scalars in `get_mask_sizes` (#40907)

Cyrilvallez committed 102 days ago

Verified 0c1839d6

Harmonize CacheLayer names (#40892)

Cyrilvallez committed 102 days ago

Verified 3688a977

[cache] Merge static sliding and static chunked layer (#40893)

Cyrilvallez committed 102 days ago

Verified 087775d1

Fix flaky `Gemma3nAudioFeatureExtractionTest::test_dither` (#40902)

ydshieh committed 102 days ago

Verified 1aff033e

Fix getter regression (#40824)

molbap committed 102 days ago

Verified 65adc3aa

Older