Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Oct 17, 2025

• Fix offload false (#2054)

  Show description for d022280
  LRL2-ModelCloud

  authored
  d022280

  Copy full SHA for d022280

• Stream quantized tensors (#2053)

  Show description for 1c26fdb
  Qubitium

  authored
  1c26fdb

  Copy full SHA for 1c26fdb

• XIELUActivation will use some weights when activation init, so can't use init_empty_weights (#2052)

  LRL2-ModelCloud

  authored
  fe065c2

  Copy full SHA for fe065c2

• use original process_batch (#2051)

  Show description for 9448e2f
  ZX-ModelCloud

  authored
  9448e2f

  Copy full SHA for 9448e2f

• Fix packing module scaling due to incorrect shared locking (#2050)

  Show description for 1c4eb3f
  Qubitium

  authored
  1c4eb3f

  Copy full SHA for 1c4eb3f

Commits on Oct 16, 2025

• [FIX] awq_moe (#2047)

  Show description for 033b406
  ZX-ModelCloud

  authored
  033b406

  Copy full SHA for 033b406

• Scores (#2049)

  Show description for aa085fb
  Qubitium

  authored
  aa085fb

  Copy full SHA for aa085fb

• Remove broken streams (#2048)

  Show description for ce6eed2
  Qubitium

  authored
  ce6eed2

  Copy full SHA for ce6eed2

• optimize eora for multi-gpu and memory usage (#2046)

  Show description for 1a4f84b
  Qubitium

  authored
  1a4f84b

  Copy full SHA for 1a4f84b

• not log token numbers if not support batch (#2044)

  LRL2-ModelCloud

  authored
  9ed4c97

  Copy full SHA for 9ed4c97

• Memory fix (#2045)

  Show description for 8df8096
  Qubitium

  authored
  8df8096

  Copy full SHA for 8df8096

• Use spinner (#2042)

  Show description for 2b9bb52
  Qubitium

  authored
  2b9bb52

  Copy full SHA for 2b9bb52

• [FIX] unittest test_packable / test_multi_gpu_inference / test_parameter_count (#2043)

  Show description for df684ef
  ZX-ModelCloud

  authored
  df684ef

  Copy full SHA for df684ef

• fix dynamic (#2041)

  LRL2-ModelCloud

  authored
  4948431

  Copy full SHA for 4948431

• Fix qwen3 moe (#2040)

  Show description for 567423c
  Qubitium

  authored
  567423c

  Copy full SHA for 567423c

Commits on Oct 15, 2025

• Tests (#2039)

  Show description for d2aba19
  Qubitium

  authored
  d2aba19

  Copy full SHA for d2aba19

• Use replicate (#2038)

  Show description for 9f95d48
  Qubitium

  authored
  9f95d48

  Copy full SHA for 9f95d48

• safety check before cpp call (#2037)

  Qubitium

  authored
  b72dd74

  Copy full SHA for b72dd74

• Add Tensor Parallel optimized weight processor (#2036)

  Show description for 26aed96
  Qubitium

  authored
  26aed96

  Copy full SHA for 26aed96

• remove logger board (#2035)

  Show description for 6fcc8cc
  Qubitium

  authored
  6fcc8cc

  Copy full SHA for 6fcc8cc

• Eora fix (#2033)

  Show description for c2457f7
  Qubitium

  authored
  c2457f7

  Copy full SHA for c2457f7

• Fix attn mask ci test (#2034)

  Show description for 6f30239
  

  Qubitium

  and

  LRL2-ModelCloud

  authored
  6f30239

  Copy full SHA for 6f30239

• [FIX] unit test and Adapted to mlx-lm 0.28.2 (#2031)

  Show description for c9fc9b2
  ZX-ModelCloud

  authored
  c9fc9b2

  Copy full SHA for c9fc9b2

• fix test_ppl (#2030)

  LRL2-ModelCloud

  authored
  52b41b5

  Copy full SHA for 52b41b5

• fix test_qzero_offsets (#2029)

  LRL2-ModelCloud

  authored
  f07f7cc

  Copy full SHA for f07f7cc

• use GSM8K_COT (#2028)

  LRL2-ModelCloud

  authored
  bdd08f5

  Copy full SHA for bdd08f5

• Make sure packer is part of dist (#2026)

  Show description for 42d0e3e
  Qubitium

  authored
  42d0e3e

  Copy full SHA for 42d0e3e

• make act_group_aware default true (#2025)

  Show description for 8a6fd21
  Qubitium

  authored
  8a6fd21

  Copy full SHA for 8a6fd21

• update scores (#2024)

  Show description for 9db2d10
  Qubitium

  authored
  9db2d10

  Copy full SHA for 9db2d10

• Fix pack block (#2023)

  Show description for 03fe86e
  Qubitium

  authored
  03fe86e

  Copy full SHA for 03fe86e

• replicate still unsafe to use (#2022)

  Show description for f7bfbbc
  Qubitium

  authored
  f7bfbbc

  Copy full SHA for f7bfbbc

• Replicate (#2021)

  Show description for 32cfacf
  Qubitium

  authored
  32cfacf

  Copy full SHA for 32cfacf

• Fix marlin kernel compiler warnings (#2020)

  Show description for da95aa5
  Qubitium

  authored
  da95aa5

  Copy full SHA for da95aa5

Commits on Oct 14, 2025

• Fix multi gpu regression (#2019)

  Show description for 208c136
  Qubitium

  authored
  208c136

  Copy full SHA for 208c136

• [FIX] test with transformers/optimum/peft (#2018)

  Show description for 4a39b3c
  ZX-ModelCloud

  authored
  4a39b3c

  Copy full SHA for 4a39b3c

Pagination

Previous

Next