Commits

Breadcrumbs

History for

onmain

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

  View code at this pointBrowse 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

  View code at this pointBrowse 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

  View code at this pointBrowse 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

  View code at this pointBrowse 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

  View code at this pointBrowse repository at this point

• Skip test_comm for nvfp4 (#4210)

  Show description for 3e6c496
  jerryzh168

  authored
  3e6c496View commit details

  Copy full SHA for 3e6c496

  View code at this pointBrowse repository at this point

Commits on Mar 31, 2026

• fix ruff on main branch (#4207)

  Show description for 79159f2
  vkuzo

  authored
  79159f2View commit details

  Copy full SHA for 79159f2

  View code at this pointBrowse 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

  View code at this pointBrowse repository at this point

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

  iamzainhuda

  authored
  3ad1067View commit details

  Copy full SHA for 3ad1067

  View code at this pointBrowse repository at this point

Commits on Mar 30, 2026

• add rope fusion detection tests (#4183)

  Show description for f11eff8
  howardzhang-cv

  authored
  f11eff8View commit details

  Copy full SHA for f11eff8

  View code at this pointBrowse 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

  View code at this pointBrowse 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

  View code at this pointBrowse 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

  View code at this pointBrowse 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

  View code at this pointBrowse 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

  View code at this pointBrowse repository at this point

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

  danielvegamyhre

  authored
  d17c61bView commit details

  Copy full SHA for d17c61b

  View code at this pointBrowse 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

  View code at this pointBrowse 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

  View code at this pointBrowse repository at this point

• Fix rocm CI (#4167)

  Show description for 6a2f643
  jerryzh168

  authored
  6a2f643View commit details

  Copy full SHA for 6a2f643

  View code at this pointBrowse repository at this point

• Move bitpacking.py to prototype and add uintx_utils.py (#4152)

  Show description for a927712
  jerryzh168

  authored
  a927712View commit details

  Copy full SHA for a927712

  View code at this pointBrowse repository at this point

• Skip test_fsdp2 if PyTorch version is 2.11.0 or higher (#4168)

  Show description for 9ea1e67
  

  Jesse Cai

  authored
  9ea1e67View commit details

  Copy full SHA for 9ea1e67

  View code at this pointBrowse repository at this point

• [reland][xpu] INT8 quantization on Intel XPU (#3782)

  Show description for 3330d29
  

  zxd1997066

  and

  liangan1

  authored
  3330d29View commit details

  Copy full SHA for 3330d29

  View code at this pointBrowse repository at this point

Commits on Mar 24, 2026

• Fix test_sparse_api failures for builds without hipSPARSELt (#4125) (#4125)

  Show description for ac0b820
  

  Benji Beck

  authored
  ac0b820View commit details

  Copy full SHA for ac0b820

  View code at this pointBrowse repository at this point

• Delete deprecated Int4CPULayout and related code (#4149)

  Show description for 1b2682d
  jerryzh168

  authored
  1b2682dView commit details

  Copy full SHA for 1b2682d

  View code at this pointBrowse repository at this point

• Delete deprecated SemiSparseLayout and related code (#4148)

  Show description for 436b2e9
  jerryzh168

  authored
  436b2e9View commit details

  Copy full SHA for 436b2e9

  View code at this pointBrowse repository at this point

• Delete deprecated UintxLayout and related code (#4145)

  Show description for abe4f57
  jerryzh168

  authored
  abe4f57View commit details

  Copy full SHA for abe4f57

  View code at this pointBrowse repository at this point

• [X86] Re-enable DA8W4 path on X86 CPU (#4033)

  Show description for 7825d99
  

  Xia-Weiwen

  and

  Copilot

  authored
  7825d99View commit details

  Copy full SHA for 7825d99

  View code at this pointBrowse repository at this point

• Delete deprecated GemlitePackedLayout and GemliteUIntXWeightOnlyConfig (#4144)

  Show description for fe98658
  jerryzh168

  authored
  fe98658View commit details

  Copy full SHA for fe98658

  View code at this pointBrowse repository at this point

• Delete deprecated BlockSparseLayout and related code (#4143)

  Show description for ea9cece
  jerryzh168

  authored
  ea9ceceView commit details

  Copy full SHA for ea9cece

  View code at this pointBrowse repository at this point

Commits on Mar 23, 2026

• [moe training] Optimize FP8 MoE backward pass: fused colwise kernel + AMD tuning (#4069)

  Show description for 34322b5
  

  4 peopleauthored
  34322b5View commit details

  Copy full SHA for 34322b5

  View code at this pointBrowse repository at this point

• Delete deprecated Float8Layout and Float8AQTTensorImpl (#4127)

  Show description for d5894a7
  jerryzh168

  authored
  d5894a7View commit details

  Copy full SHA for d5894a7

  View code at this pointBrowse repository at this point

Commits on Mar 22, 2026

• Delete deprecated CutlassSemiSparseLayout and related code (#4126)

  Show description for 4cef023
  jerryzh168

  authored
  4cef023View commit details

  Copy full SHA for 4cef023

  View code at this pointBrowse repository at this point

• Delete deprecated autoquant v1 and all references (#4122)

  Show description for 5812c35
  jerryzh168

  authored
  5812c35View commit details

  Copy full SHA for 5812c35

  View code at this pointBrowse repository at this point

Commits on Mar 20, 2026

• Add CuTeDSL kernel for 3D tensor quantization to MXFP8 (#4090)

  alexsamardzic

  authored
  d7509a6View commit details

  Copy full SHA for d7509a6

  View code at this pointBrowse repository at this point

• [ROCm] Fix ROCm CI failures (#4061)

  Show description for 51f1116
  brucechanglongxu

  authored
  51f1116View commit details

  Copy full SHA for 51f1116

  View code at this pointBrowse repository at this point