huggingface/transformers

Pull Requests Commits

Fix model parallel device mismatch in `create_bidirectional_sliding_window_mask` (#47560)

abcgco committed 27 days ago

Verified af711556

CI: use a single function for GH calls (#47474)

tarekziade committed 27 days ago

Verified d0e91c7f

Kernels and loaders robustification (#47334)

IlyasMoutawwakil committed 27 days ago

Verified ebea912f

Fix A.X-K2 fp8 modules_to_not_convert normalization for the gated-norm MLP (#47578)

kmswin1 committed 27 days ago

Verified ff24c90c

Remove redundant guarding for distributed (#47570)

3outeille committed 28 days ago

Verified dff4572d

Add FSDP plans to all models (#47165)

3outeille committed 28 days ago

Verified 3e9d3e50

Fix modular for mamba packages (#47494)

Cyrilvallez committed 28 days ago

Verified 77f69441

Remove deprecated conversion in Kimi (#47581)

zucchini-nlp committed 28 days ago

Verified 44752c8d

CI: use transformers-ci daily workflow with OTEL (#47360)

tarekziade committed 28 days ago

Verified aac2512e

Fix slow tensor path in _check_special_mm_tokens (#47580)

guan404ming committed 28 days ago

Verified f3f0c8b7

CI: let's run integration failure cron at 10pm (#47537)

tarekziade committed 28 days ago

Verified 4a311bbb

Better and more extensive tests for RoPE (#46912)

zucchini-nlp committed 28 days ago

Verified 5b6f170a

Fix mamba2 family decode and simplify all reshape ops (#47569)

Cyrilvallez committed 28 days ago

Verified 9a223f15

Speed up image preprocessing for vision-language models (#47453)

labAxiaoming committed 28 days ago

Verified 41db33d8

[DiffusionGemma] Cast the decoder padding mask to bool (#47295)

kashif committed 28 days ago

Verified 4ad6ba18

Fix NemotronH: Register `"mlp"` in the cache layer-type mappings (#47535)

qgallouedec committed 28 days ago

Verified 016abe1a

Fix `config.per_layer_config[layer_type]` lookup speed (#47539)

hmellor committed 28 days ago

Verified 142f01fb

General maintenance (#47517)

zucchini-nlp committed 28 days ago

Verified 076ad909

:rotating_light: Processors update the rest (#46556)

zucchini-nlp committed 28 days ago

Verified 254e4b6e

fixed the benchmark script with DistributedConfig (#47568)

tarekziade committed 28 days ago

Verified 5f8521c0

Fix TP inference for tied embedding (#47503)

3outeille committed 29 days ago

Verified 279e6efc

Add AXK2 from SKT (#47528)

vasqu committed 31 days ago

Verified b6d5084f

Fix `causal_conv1d_fn` positional `activation` colliding with hub kernel's `seq_idx` (#47527)

qgallouedec committed 31 days ago

Verified 77a78c9f

Fix failing tests for mimo_v2_flash (#47284)

kaixuanliu committed 31 days ago

Verified ffec245b

Refactor all linear attention models to latest best standards for convolution (#47452)

Cyrilvallez committed 31 days ago

Verified a79421ac

Make tokenization_mistral_common importable without mistral_common installed (#47397)

juliendenize committed 31 days ago

Verified 98a1e7d8

Use --flake-runs=1 in check_bad_commit.py for PR comment CI (#47522)

ydshieh committed 31 days ago

Verified 5b32a41f

Deprecate the old response_schema (#47320)

Rocketknight1 committed 31 days ago

Verified 77bcb281

fix: add pickle support to _LazyConfigMapping for spawn multiprocessing (#46026)

kfojcik-intel committed 31 days ago

Verified f07253a9

Fix potential ReDoS by escaping tokenizer filename used as regex pattern (#47498)

hameedibrh committed 31 days ago

Verified 7e0ff853

Newer Older