Commits

Branch selector

main

User selector

All users

All time

Commit history

Commits on Apr 1, 2026

• [ROCm] Enable FP8 QAT test on MI300/MI350 (#4132)

  brucechanglongxu

  authored
  6807454View commit details

  Copy full SHA for 6807454

  Browse repository at this point

• add hadamard option to low precision attention api (#4194)

  Show description for 168a9e2
  howardzhang-cv

  authored
  168a9e2View commit details

  Copy full SHA for 168a9e2

  Browse repository at this point

• [CPU] Rename Int4WeightOnlyOpaqueTensorConfig to PrototypeInt4WeightOnlyConfig (#4205)

  Show description for 432a867
  

  Xia-Weiwen

  and

  Copilot

  authored
  432a867View commit details

  Copy full SHA for 432a867

  Browse repository at this point

• [CPU] Re-enable the previously failing test_fp8_q_attention_block UT (#4212)

  Show description for 8ed52c0
  Yuxingwang-intel

  authored
  8ed52c0View commit details

  Copy full SHA for 8ed52c0

  Browse repository at this point

• Delete deprecated TensorCoreTiledLayout and related code (#4153)

  Show description for 3e0da46
  jerryzh168

  authored
  3e0da46View commit details

  Copy full SHA for 3e0da46

  Browse repository at this point

• Fix NF4Tensor narrow() crash with recent PyTorch nightly (#4208)

  Show description for 1dbe5bf
  

  vkuzo

  and

  claude

  authored
  1dbe5bfView commit details

  Copy full SHA for 1dbe5bf

  Browse repository at this point

• Bump version from 0.17.0 to 0.18.0 (#4172)

  

  Jesse Cai

  authored
  e41fb14View commit details

  Copy full SHA for e41fb14

  Browse repository at this point

• [CPU] Update qsdpa lowering to return only the output node (#4162)

  Show description for 62212e4
  Yuxingwang-intel

  authored
  62212e4View commit details

  Copy full SHA for 62212e4

  Browse repository at this point

• Skip test_comm for nvfp4 (#4210)

  Show description for 3e6c496
  jerryzh168

  authored
  3e6c496View commit details

  Copy full SHA for 3e6c496

  Browse repository at this point

Commits on Mar 31, 2026

• Add llms.txt to docs site for LLM crawler discoverability (#4196)

  Show description for 5b87fea
  supriyar

  authored
  5b87feaView commit details

  Copy full SHA for 5b87fea

  Browse repository at this point

• Add CLAUDE.md with AI agent instructions and quick reference (#4195)

  Show description for 7c1b138
  supriyar

  authored
  7c1b138View commit details

  Copy full SHA for 7c1b138

  Browse repository at this point

• fix ruff on main branch (#4207)

  Show description for 79159f2
  vkuzo

  authored
  79159f2View commit details

  Copy full SHA for 79159f2

  Browse repository at this point

• Create recipe for flux2pro running on AMD (#4200)

  Show description for 7dd8be5
  hualazimi0425

  authored
  7dd8be5View commit details

  Copy full SHA for 7dd8be5

  Browse repository at this point

• install torchvision to resolve ci issue (#4203)

  danielvegamyhre

  authored
  da257b5View commit details

  Copy full SHA for da257b5

  Browse repository at this point

• add torch.compile test for Float8BlockwiseLinear (#4187)

  iamzainhuda

  authored
  3ad1067View commit details

  Copy full SHA for 3ad1067

  Browse repository at this point

Commits on Mar 30, 2026

• [xpu][test] Add weekly test for Intel GPU (#4190)

  Show description for f26135c
  zxd1997066

  authored
  f26135cView commit details

  Copy full SHA for f26135c

  Browse repository at this point

• Bump version from 0.17.0 to 0.18.0 (#4202)

  

  Jesse Cai

  authored
  0feb93cView commit details

  Copy full SHA for 0feb93c

  Browse repository at this point

• add rope fusion detection tests (#4183)

  Show description for f11eff8
  howardzhang-cv

  authored
  f11eff8View commit details

  Copy full SHA for f11eff8

  Browse repository at this point

• Add CodeML'25 BibTeX (#3978)

  Show description for ce07646
  namgyu-youn

  authored
  ce07646View commit details

  Copy full SHA for ce07646

  Browse repository at this point

• [X86] intmm: Use u8s8 when only support avx512-vnni (#4103)

  Show description for d5814ae
  cyxlily

  authored
  d5814aeView commit details

  Copy full SHA for d5814ae

  Browse repository at this point

• [CPU] Add software prefetch to overlap bandwidth for scaled_embedding_bag (#4171)

  Show description for 1087d59
  yanbing-j

  authored
  1087d59View commit details

  Copy full SHA for 1087d59

  Browse repository at this point

Commits on Mar 27, 2026

• [mxfp8 moe training] register sharding rules for cutedsl 2d quant kernel (#4178)

  danielvegamyhre

  authored
  f6f29f6View commit details

  Copy full SHA for f6f29f6

  Browse repository at this point

• fix undefined values for tail elements in act quant kernels (#4186)

  iamzainhuda

  authored
  a6be48fView commit details

  Copy full SHA for a6be48f

  Browse repository at this point

• Add better debug print for failed prepare (#3950)

  Show description for 9051a2f
  JakeStevens

  authored
  9051a2fView commit details

  Copy full SHA for 9051a2f

  Browse repository at this point

Commits on Mar 26, 2026

• [mxfp8 moe training] remove unused block_size arg (#4177)

  danielvegamyhre

  authored
  96a9cdfView commit details

  Copy full SHA for 96a9cdf

  Browse repository at this point

• [mxfp8 moe training] refactor autograd func (#4176)

  danielvegamyhre

  authored
  efbcb0eView commit details

  Copy full SHA for efbcb0e

  Browse repository at this point

• Add WAN style rope detection pattern for low precision attention API (#4173)

  Show description for 6f56403
  howardzhang-cv

  authored
  6f56403View commit details

  Copy full SHA for 6f56403

  Browse repository at this point

• remove per kernel fp8 triton quantized benchmarks (#4166)

  iamzainhuda

  authored
  2c4ecb0View commit details

  Copy full SHA for 2c4ecb0

  Browse repository at this point

• Add embedding op support for Float8Tensor with PerGroup quantization (#4175)

  Show description for 4611835
  jerryzh168

  authored
  4611835View commit details

  Copy full SHA for 4611835

  Browse repository at this point

• Add PerGroup granularity for Float8WeightOnly config (#4174)

  Show description for 4826062
  jerryzh168

  authored
  4826062View commit details

  Copy full SHA for 4826062

  Browse repository at this point

Commits on Mar 25, 2026

• [mxfp8 training] add cutedsl kernel for mxfp8 quantation along dim0 (#4156)

  Show description for 02105d4
  danielvegamyhre

  authored
  02105d4View commit details

  Copy full SHA for 02105d4

  Browse repository at this point

• clean up unused rocm references in test_training.py (#4170)

  danielvegamyhre

  authored
  d17c61bView commit details

  Copy full SHA for d17c61b

  Browse repository at this point

• Remove tensor parallel test for v1 of Int8DynamicActivationInt8WeightConfig (#4169)

  Show description for 136cacb
  jerryzh168

  authored
  136cacbView commit details

  Copy full SHA for 136cacb

  Browse repository at this point

• [xpu][test] Skip WIP config for Intel GPU in test_safetensors_support.py and test_x86inductor_fusion.py (#4049)

  Show description for 8fca033
  zxd1997066

  authored
  8fca033View commit details

  Copy full SHA for 8fca033

  Browse repository at this point

• Fix rocm CI (#4167)

  Show description for 6a2f643
  jerryzh168

  authored
  6a2f643View commit details

  Copy full SHA for 6a2f643

  Browse repository at this point