Commits

Branch selector

 main

User selector

mgoin

Datepicker

All time

Commit History

Commits on May 27, 2025

• [Bugfix] Mistral tool calling when content is list (#18729)

  Show description for 5873877
  mgoin

  authored
  5873877

  Copy full SHA for 5873877

Commits on May 25, 2025

• Speed up the `kernels/quantization/` tests (#18669)

  Show description for 6393454
  mgoin

  authored
  6393454

  Copy full SHA for 6393454

Commits on May 23, 2025

• [CI] Enable test_initialization to run on V1 (#16736)

  Show description for 0ddf88e
  mgoin

  authored
  0ddf88e

  Copy full SHA for 0ddf88e

• [Docs] Change mkdocs to not use directory urls (#18622)

  Show description for 9520a98
  mgoin

  authored
  9520a98

  Copy full SHA for 9520a98

• [Doc] Update quickstart and install for cu128 using `--torch-backend=auto` (#18505)

  Show description for 54af915
  mgoin

  authored
  54af915

  Copy full SHA for 54af915

Commits on May 21, 2025

• [TPU] Re-enable the Pallas MoE kernel (#18025)

  Show description for 3b17ea2
  mgoin

  authored
  3b17ea2

  Copy full SHA for 3b17ea2

Commits on May 20, 2025

• [Minor] Rename quantization nvfp4 to modelopt_fp4 (#18356)

  Show description for f4a8a37
  mgoin

  authored
  f4a8a37

  Copy full SHA for f4a8a37

Commits on May 17, 2025

• Update Dockerfile to build for Blackwell (#18095)

  mgoin

  authored
  dcfe952

  Copy full SHA for dcfe952

• [V1][P/D] Local attention optimization for NIXL (#18170)

  Show description for fd195b1
  mgoin

  authored
  fd195b1

  Copy full SHA for fd195b1

Commits on May 14, 2025

• [V1] Support multiple kv connectors (#17564)

  Show description for 2142035
  

  mgoin

  and

  njhill

  authored
  2142035

  Copy full SHA for 2142035

• [Bugfix][V1] Fix FlashInfer V1 backend using the wrong VllmConfig (#18086)

  mgoin

  authored
  12e6c0b

  Copy full SHA for 12e6c0b

• [Bugfix] Fix FP8 Marlin MoE and enable for compressed-tensors models (#18026)

  Show description for 9a2a635
  mgoin

  authored
  9a2a635

  Copy full SHA for 9a2a635

Commits on May 13, 2025

• [Bugfix] Fix marlin moe fallback logic for llama4 (#18042)

  Show description for ea6ae8c
  mgoin

  authored
  ea6ae8c

  Copy full SHA for ea6ae8c

• [Bugfix] Fixes for new marlin moe usage (#18017)

  Show description for 1df491c
  mgoin

  authored
  1df491c

  Copy full SHA for 1df491c

• Use NVFP4 Marlin for CompressedTensorsW4A16Fp4 (#18000)

  Show description for 307939f
  

  mgoin

  and

  dsikka

  authored
  307939f

  Copy full SHA for 307939f

Commits on May 12, 2025

• Fix FBGEMM integration (#18002)

  Show description for f065de4
  mgoin

  authored
  f065de4

  Copy full SHA for f065de4

Commits on May 9, 2025

• Update CT WNA16MarlinMoE integration (#16666)

  Show description for 22481fb
  mgoin

  authored
  22481fb

  Copy full SHA for 22481fb

• Revert "[BugFix][AMD] Compatible patch for latest AITER(05/07/2025)" (#17910)

  mgoin

  authored
  85b72cb

  Copy full SHA for 85b72cb

Commits on May 8, 2025

• Fix noisy warning for uncalibrated q_scale/p_scale (#17414)

  Show description for 4f605a6
  mgoin

  authored
  4f605a6

  Copy full SHA for 4f605a6

• [CI] Prune down lm-eval small tests (#17012)

  Show description for 8342e3a
  mgoin

  authored
  8342e3a

  Copy full SHA for 8342e3a

Commits on May 7, 2025

• Replace lm-eval bash script with pytest and use enforce_eager for faster CI (#17717)

  Show description for 950b711
  mgoin

  authored
  950b711

  Copy full SHA for 950b711

• [TPU] Add kernel test for moe_pallas (#17496)

  Show description for e50a1f1
  mgoin

  authored
  e50a1f1

  Copy full SHA for e50a1f1

• Removed unused marlin cuda code (#17684)

  Show description for a17cef7
  mgoin

  authored
  a17cef7

  Copy full SHA for a17cef7

Commits on May 6, 2025

• [V1] Enable TPU V1 backend by default (#17673)

  Show description for d419aa5
  mgoin

  authored
  d419aa5

  Copy full SHA for d419aa5

• Update nm to rht in doc links + refine fp8 doc (#17678)

  Show description for 98834fe
  mgoin

  authored
  98834fe

  Copy full SHA for 98834fe

Commits on May 2, 2025

• Support W8A8 INT8 MoE for compressed-tensors (#16745)

  Show description for 868c546
  mgoin

  authored
  868c546

  Copy full SHA for 868c546

• Fix PixtralHF missing spatial_merge_size (#17571)

  Show description for f192ca9
  mgoin

  authored
  f192ca9

  Copy full SHA for f192ca9

• Check if bitblas is installed during support check (#17572)

  Show description for b4003d1
  mgoin

  authored
  b4003d1

  Copy full SHA for b4003d1

• [CI] Actually run tests/kv_transfer/test_disagg.py in CI (#17555)

  Show description for 292fc59
  mgoin

  authored
  292fc59

  Copy full SHA for 292fc59

• [Bugfix] Disable gptq_bitblas for <SM80 to fix GPTQ on V100/T4 (#17541)

  Show description for 24aebae
  mgoin

  authored
  24aebae

  Copy full SHA for 24aebae

Commits on May 1, 2025

• [CI][Bugfix] Fix failing V1 Test due to missing 'cache_salt' arg (#17500)

  Show description for aa4502e
  mgoin

  authored
  aa4502e

  Copy full SHA for aa4502e

• [CI][TPU] Skip structured outputs+spec decode tests on TPU (#17510)

  Show description for 17b4d85
  mgoin

  authored
  17b4d85

  Copy full SHA for 17b4d85

Commits on Apr 30, 2025

• [Docs] Update optimization.md doc (#17482)

  Show description for 0b7e701
  mgoin

  authored
  0b7e701

  Copy full SHA for 0b7e701

• Support LoRA for Mistral3 (#17428)

  Show description for a44c4f1
  mgoin

  authored
  a44c4f1

  Copy full SHA for a44c4f1

Commits on Apr 28, 2025

• [Model] Add tuned triton fused_moe configs for Qwen3Moe (#17328)

  Show description for 8fc88d6
  mgoin

  authored
  8fc88d6

  Copy full SHA for 8fc88d6

Pagination

Previous

Next