Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on Feb 12, 2026

• [Bugfix] Fix Sparse24 Compressed Tensors models (#33446)

  Show description for e9cd691
  

  kylesayrs

  and

  mgoin

  authored
  e9cd691

  Copy full SHA for e9cd691

Commits on Feb 11, 2026

• [Bugfix][CPU] Fix llama4 inference on CPU (#34321)

  Show description for 05339a7
  bigPYJ1151

  authored
  05339a7

  Copy full SHA for 05339a7

• [CPU] Enable FP16 (Half dtype) support for s390x (#34116)

  Show description for d1b837f
  R3hankhan123

  authored
  d1b837f

  Copy full SHA for d1b837f

• [Kernel] Apply 256bit LDG/STG To Activation Kernels (#33022)

  Show description for 1485396
  

  AstroVoyager7

  and

  yewentao256

  authored
  1485396

  Copy full SHA for 1485396

• [Bugfix][DeepSeek-V3.2] fix fp8 kvcache type cast (#33884)

  Show description for 5ee5c86
  kebe7jun

  authored
  5ee5c86

  Copy full SHA for 5ee5c86

Commits on Feb 10, 2026

• [Perf][Kernel] Add faster topKperRow decode kernel for DeepSeek-V3.2 sparse attention (#33680)

  Show description for afdce12
  

  LopezCastroRoberto

  and

  claude

  authored
  afdce12

  Copy full SHA for afdce12

Commits on Feb 9, 2026

• [Fix] [CPU Backend] : Prepack weights for w8a8 oneDNN matmul (#33901)

  Show description for caad9f1
  nikhil-arm

  authored
  caad9f1

  Copy full SHA for caad9f1

• fix(cpu): fix mla_decode compilation on x86 without AVX512 (#34052)

  Show description for 5a5c435
  

  ihb2032

  and

  root

  authored
  5a5c435

  Copy full SHA for 5a5c435

Commits on Feb 7, 2026

• Perf tuning and expansion of cases covered for wvSplitKrc (#33493)

  Show description for ed17f54
  amd-hhashemi

  authored
  ed17f54

  Copy full SHA for ed17f54

• [Kernel] Add enable_sm120_or_later for SM121 (DGX Spark) CUTLASS support (#33517)

  Show description for bc32444
  Code4me2

  authored
  bc32444

  Copy full SHA for bc32444

Commits on Feb 6, 2026

• [Refactor] Remove align block size logic in `moe_permute` (#33449)

  Show description for 77c09e1
  yewentao256

  authored
  77c09e1

  Copy full SHA for 77c09e1

• [cpu][performance] CPU Paged Attention NEON BFMMLA BF16 Implementation (#32263)

  Show description for 1363e3d
  gassan-arm

  authored
  1363e3d

  Copy full SHA for 1363e3d

• [CPU] Add BF16 Kernel type for s390x (#33788)

  Show description for ac04dd3
  R3hankhan123

  authored
  ac04dd3

  Copy full SHA for ac04dd3

Commits on Feb 5, 2026

• Adds padding and perf improvements to wvSplitK_fp8 (#33527)

  Show description for d5c4800
  amd-hhashemi

  authored
  d5c4800

  Copy full SHA for d5c4800

Commits on Feb 4, 2026

• [CPU] Split attention dispatch by head_dim alignment (#32161)

  Show description for 4dffc5e
  R3hankhan123

  authored
  4dffc5e

  Copy full SHA for 4dffc5e

Commits on Feb 3, 2026

• [Feature][CPU Backend]: Optimize ARM vectorization backend (#30329)

  Show description for e69c990
  Radu2k

  authored
  e69c990

  Copy full SHA for e69c990

Commits on Feb 2, 2026

• fix cutlass_3x_gemm_fp8_blockwise on sm103a (#32224)

  Show description for 089cd4f
  

  IwakuraRein

  and

  pavanimajety

  authored
  089cd4f

  Copy full SHA for 089cd4f

• [Feature][Core] Support Fabric detection to adapt the MNNVL protocol for the GB series (#33540)

  Show description for 528e9b1
  

  3 peopleauthored
  528e9b1

  Copy full SHA for 528e9b1

Commits on Jan 31, 2026

• [Bugfix]: Fix display errors in TORCH_CHECK messages (#32942)

  Show description for fedf643
  

  lingebeng

  and

  yewentao256

  authored
  fedf643

  Copy full SHA for fedf643

Commits on Jan 29, 2026

• [Bugfix][CPU] Fix thread num for shared memory communication (#33317)

  Show description for 8311f08
  

  bigPYJ1151

  and

  gemini-code-assist[bot]

  authored
  8311f08

  Copy full SHA for 8311f08

• [Doc]: fixing multiple typos in diverse files (#33256)

  Show description for 31b25f6
  

  didier-durand

  and

  gemini-code-assist[bot]

  authored
  31b25f6

  Copy full SHA for 31b25f6

Commits on Jan 28, 2026

• [Perf] Optimize `moe_permute` for CUTLASS FP8 (#32892)

  Show description for c4e744d
  yewentao256

  authored
  c4e744d

  Copy full SHA for c4e744d

• [Quantization][Deprecation] Remove Marlin 24 (#32688)

  Show description for af9b69f
  

  3 peopleauthored
  af9b69f

  Copy full SHA for af9b69f

• Add flake8-implicit-str-concat rules to Ruff (#33191)

  Show description for 2eb673a
  hmellor

  authored
  2eb673a

  Copy full SHA for 2eb673a

Commits on Jan 27, 2026

• [AMD][Kernel][BugFix] Use correct scale in concat_and_cache_ds_mla_kernel when on gfx942 (#32976)

  Show description for 58996f3
  

  rasmith

  and

  Randall Smith

  authored
  58996f3

  Copy full SHA for 58996f3

Commits on Jan 26, 2026

• [fix] CPUDNNLGEMMHandler pointer baked into inductor artifact (#32913)

  Show description for 58a05b0
  dolpm

  authored
  58a05b0

  Copy full SHA for 58a05b0

Commits on Jan 25, 2026

• [Perf][Kernel] Optimize FP4 quantization kernels (SM100F) (#32520)

  Show description for fcb9df9
  LopezCastroRoberto

  authored
  fcb9df9

  Copy full SHA for fcb9df9

Commits on Jan 23, 2026

• [Refactor] Rename `gptq_marlin` to `marlin` to match MoE (#32952)

  Show description for 4561f13
  mgoin

  authored
  4561f13

  Copy full SHA for 4561f13

• [CI][AMD][BugFix] Update wvSplitK (and other skinny_gemm wrappers) to ensure tensors passed will be made contiguous for the kernel (#32831)

  Show description for 6cc6d92
  

  rasmith

  and

  Randall Smith

  authored
  6cc6d92

  Copy full SHA for 6cc6d92

• [CI/Build][CPU] Fix failed pooling tests and macos smoke test (#32907)

  Show description for 5da4c7d
  

  bigPYJ1151

  and

  gemini-code-assist[bot]

  authored
  5da4c7d

  Copy full SHA for 5da4c7d

Commits on Jan 22, 2026

• Add llmcompressor fp8 kv-cache quant (per-tensor and per-attn_head) (#30141)

  Show description for 44f08af
  eldarkurtic

  authored
  44f08af

  Copy full SHA for 44f08af

• [CPU Backend] [Perf] Accelerate tensor-parallel/data-parallel inference across NUMA domains on Arm (#32792)

  Show description for 744ef30
  fadara01

  authored
  744ef30

  Copy full SHA for 744ef30

• OffloadingConnector: Support kernel_block_size != block_size (#30692)

  Show description for 421012b
  orozery

  authored
  421012b

  Copy full SHA for 421012b

Commits on Jan 21, 2026

• [Kernel] Add topk_sigmoid kernel (#31246)

  Show description for 63227ac
  xyang16

  authored
  63227ac

  Copy full SHA for 63227ac

• [Quantization][Deprecation] Deprecate HQQ (#32681)

  Show description for 85f55c9
  

  robertgshaw2-redhat

  and

  Robert Shaw

  authored
  85f55c9

  Copy full SHA for 85f55c9