NielsRogge/transformers

Pull Requests Commits

Switch to use A10 progressively (#38936)

ydshieh committed 1 year ago

Verified 3d34b921

Fix more flaky `test_initialization` (#38932)

ydshieh committed 1 year ago

Verified b8059e1f

Correctly raise error for awq quantization (#38945)

Cyrilvallez committed 1 year ago

Verified 5ee60f97

Pin PyTorch extras for AMD containers (#38941)

ahadnagy committed 1 year ago

Verified 8ac2d753

Add kwargs for timm.create_model in TimmWrapper (#38860)

qubvel committed 1 year ago

Verified 9120567b

[static cache] fix device map per layer in VLMs (#38488)

zucchini-nlp committed 1 year ago

Verified ff95974b

Remove `ALL_LAYERNORM_LAYERS` (#38922)

Cyrilvallez committed 1 year ago

Verified aa42987c

add pytorch-xpu Dockerfile (#38875)

yao-matrix committed 1 year ago

Verified 38a9b707

Modernbert fixes (#38912)

remi-or committed 1 year ago

Verified 9bcdd5cd

Skip some tests for now (#38931)

ydshieh committed 1 year ago

Verified 31d30b72

Remove deprecated classes in modeling_utils.py (#38919)

Cyrilvallez committed 1 year ago

Verified 0725cd69

feat: add flexible Liger Kernel configuration to TrainingArguments (#38911)

hamza-hcompany committed 1 year ago

Verified 797860c6

Allow make-fixup on main branch, albeit slowly (#38892)

Rocketknight1 committed 1 year ago

Verified 89b35be6

feat: Add granite architectures to auto tokenizer name mappings (#38802)

gabe-l-hart committed 1 year ago

Verified 9a02e760

Fix ReDOS in tokenizer digit substitution (#38844)

Rocketknight1 committed 1 year ago

Verified 54a02160

Skip sdpa tests if submodule does not support sdpa (#38907)

ivarflakstad committed 1 year ago

Verified af6120b3

Fix `FalconMambaIntegrationTests` (#38566)

ydshieh committed 1 year ago

Verified 5d26a387

align xpu's autocast behavior w/ cuda by using device agnostic torch APIs (#38284)

yao-matrix committed 1 year ago

Verified a9ce8c69

Fix unnecessary super calls (#38897)

cyyever committed 1 year ago

Verified 0a53df1a

Fix `fsmt` tests (#38904)

ydshieh committed 1 year ago

Verified b949747b

[phi-4] use mel filters from audio utils (#36966)

eustlb committed 1 year ago

Verified 11738f85

Use `raise from e` in `hub.py` utility (#37241)

Wauplin committed 1 year ago

Verified f7b21822

Add support for specifying revisions when pushing to Hub via internal Trainer call (#36852)

IsaacBreen committed 1 year ago

Verified 3756bf19

Update bamba model card (#38853)

druvdub committed 1 year ago

Verified 458e0b37

[video processor] fix slow tests (#38881)

zucchini-nlp committed 1 year ago

Verified ea013348

36978 | Fast image processor for DPT model (#37481)

samrae7 committed 1 year ago

Verified b922b22e

Docs: Add custom fine-tuning tutorial to TrOCR model page (#38847)

Ashutosh-4485 committed 1 year ago

Verified c27f628e

log: Add logging when using split_batches and per_device_train_batch_size (#38633)

KeshavSingh29 committed 1 year ago

Verified 0a289d16

[bugfix] fix ATTN_MASK_NPU device mismatch error on multi-device NPU … (#38876)

qykong committed 1 year ago

Verified c55d8063

Fix loop var naming (#38885)

Rocketknight1 committed 1 year ago

Verified 9cd7570f

Newer Older