huggingface/transformers

Pull Requests Commits

added sonicmoe a2a ep

IlyasMoutawwakil committed 24 days ago

79e0e4ef

IlyasMoutawwakil committed 24 days ago

70c7c3b9

more better errors for deepgemm

IlyasMoutawwakil committed 24 days ago

751c8b5c

better errors across deepgemm

IlyasMoutawwakil committed 24 days ago

8bd9619e

IlyasMoutawwakil committed 24 days ago

b47b30f0

add flash base test and fix the incorrect and crash issue during enab… (#46076)

sywangyi committed 24 days ago

Verified 143835f3

unnecessary change

IlyasMoutawwakil committed 24 days ago

c1adcd70

IlyasMoutawwakil committed 24 days ago

d0c4a6df

support transposed

IlyasMoutawwakil committed 24 days ago

f0e2859e

force stable abi build for now

IlyasMoutawwakil committed 24 days ago

3e76df5e

minimal changes i quantizer

IlyasMoutawwakil committed 24 days ago

619b5e9b

always floats scal inv with triton

IlyasMoutawwakil committed 24 days ago

afd1b708

remove force al kernels

IlyasMoutawwakil committed 24 days ago

2e4203fa

fix quantizer and use upstream deepgemm

IlyasMoutawwakil committed 24 days ago

78fcee66

Merge remote-tracking branch 'origin/main' into deepgemm-isolation

IlyasMoutawwakil committed 25 days ago

57081ccf

[Gemma4] Replace one-hot matmul with F.embedding in position embeddings (#46176)

Sriniketh24 committed 25 days ago

Verified bc8f70a9

fix: kosmos2.5: properly expand embeddings table (#45835)

nunq committed 25 days ago

Verified fd1d5b7d

find pytest launch error in torch 2.13.0.dev20260526 (#46252)

sywangyi committed 25 days ago

Verified f336a53d

Update a test to avoid writing to the default xet cache (#46250)

ydshieh committed 25 days ago

Verified 2d31b762

[Test][Kosmos2.5] Add XPU expectations for integration tests (#46135)

YangKai0616 committed 25 days ago

Verified 88070eff

[`Revert`] FSDP+Dtensor refactor related changes (#46246)

vasqu committed 25 days ago

Verified 295cee3e

Support FA2 flash_attn_with_kvcache for XPU continuous batching (#46028)

YangKai0616 committed 25 days ago

Verified a65bf6c9

[`Configs`] Fix layer type validation to include its mlp counterpart (#46220)

vasqu committed 25 days ago

Verified 70257e9a

Fix `num_items_in_batch` over-counting for causal LM losses (#46204)

qgallouedec committed 25 days ago

Verified 67265ef5

RF-DETR doc fixes (#46244)

merveenoyan committed 25 days ago

Verified 5adca7bf

Use `main` instead of commit SHA for now (#46241)

ydshieh committed 25 days ago

Verified 1a5a5a71

Enable push event (to main) for PR CI workflow (#46240)

ydshieh committed 26 days ago

Verified 5298d0ca

fix(hrm_text): Add XPU Expectations for tests (#46214)

kaixuanliu committed 26 days ago

Verified a6cd3b4e

Fix model parallel bugs for Gemma4 (#45817)

kaixuanliu committed 26 days ago

Verified f39b5c8b

Fix bnb 4bit/8bit quantization drop chunked tensors bug (#46210)

kaixuanliu committed 26 days ago

Verified 212863bc

Older