Commits

Branch selector

 main

User selector

All users

Datepicker

All time

Commit History

Commits on Jun 27, 2025

• fixed ring mac error (#3008)

  greenrazer

  authored
  d0a3b33

  Copy full SHA for d0a3b33

Commits on Jun 26, 2025

• Qwen3: fix quality loss due to rope freq precision (#3005)

  Show description for ab14581
  zackangelo

  authored
  ab14581

  Copy full SHA for ab14581

• Fixed Failing CI (#3007)

  Show description for 0cd4fc4
  greenrazer

  authored
  0cd4fc4

  Copy full SHA for 0cd4fc4

• fixed errors with hardswish merge (#3006)

  greenrazer

  authored
  a6e8aae

  Copy full SHA for a6e8aae

Commits on Jun 24, 2025

• candle-onnx: Implement Hard Swish operator (#2980)

  Show description for 2e5dbc7
  Michall00

  authored
  2e5dbc7

  Copy full SHA for 2e5dbc7

• Fix typos (#2958)

  omahs

  authored
  23968db

  Copy full SHA for 23968db

• candle-onnx: Implement RNN operator (#2964)

  Show description for 407c667
  BrunoSienkiewicz

  authored
  407c667

  Copy full SHA for 407c667

Commits on Jun 7, 2025

• Fix cuda memory error for Qwen3 non-quantized (#2987)

  Show description for 17313a4
  akshayballal95

  authored
  17313a4

  Copy full SHA for 17313a4

Commits on May 31, 2025

• Add Qwen3 MoE (#2934)

  Show description for 0224a74
  greenrazer

  authored
  0224a74

  Copy full SHA for 0224a74

Commits on May 30, 2025

• candle-onnx: Implement Trilu and ScatterND ops (#2952)

  Show description for cd7b877
  greenrazer

  authored
  cd7b877

  Copy full SHA for cd7b877

Commits on May 29, 2025

• feat: enhance linear algebra operations (#2972)

  Show description for 5aed817
  ssfdust

  authored
  5aed817

  Copy full SHA for 5aed817

Commits on May 28, 2025

• Add fine-tuned text classifier to xlm roberta example (#2969)

  jpe90

  authored
  1a183c9

  Copy full SHA for 1a183c9

• (hotfix) fix the doc test for indexer (#2970)

  klion26

  authored
  cac51fe

  Copy full SHA for cac51fe

Commits on May 26, 2025

• Use a tanh activation in the xlm-roberta classification head. (#2968)

  LaurentMazare

  authored
  61ddb95

  Copy full SHA for 61ddb95

Commits on May 21, 2025

• Proper support for phi-4 (#2960)

  Show description for 9a62c91
  LaurentMazare

  authored
  9a62c91

  Copy full SHA for 9a62c91

Commits on May 15, 2025

• Fixes for clippy 1.87. (#2956)

  LaurentMazare

  authored
  92106c8

  Copy full SHA for 92106c8

• Fix docs quantized qwen3 (#2955)

  Show description for 9ce4fe6
  maximizemaxwell

  authored
  9ce4fe6

  Copy full SHA for 9ce4fe6

Commits on May 14, 2025

• Olmo 2 model (#2954)

  Show description for 450a49e
  

  janimo

  and

  LaurentMazare

  authored
  450a49e

  Copy full SHA for 450a49e

• Make tensor contiguous before the repeat_kv calls to avoid strided copies (#2953)

  b0r3k

  authored
  6bd6172

  Copy full SHA for 6bd6172

Commits on May 13, 2025

• Fixed Quantized Qwen3 Model (#2951)

  Show description for 485ddf2
  nosnakeob

  authored
  485ddf2

  Copy full SHA for 485ddf2

Commits on May 10, 2025

• Add Resize to onnx ops (#2946)

  Show description for 36508a2
  greenrazer

  authored
  36508a2

  Copy full SHA for 36508a2

Commits on May 8, 2025

• Qwen3 quantized implementation (#2939)

  Show description for 3d05f5c
  

  ljt019

  and

  LaurentMazare

  authored
  3d05f5c

  Copy full SHA for 3d05f5c

Commits on May 4, 2025

• Bump cudarc to 0.16.3. (#2942)

  LaurentMazare

  authored
  637473c

  Copy full SHA for 637473c

Commits on May 3, 2025

• Indexing with max-value results in zero/no-op. (#2940)

  Show description for e27b470
  LaurentMazare

  authored
  e27b470

  Copy full SHA for e27b470

Commits on May 2, 2025

• Updating `Add qwen3` (PR 2903) to use HF weights (#2930)

  Show description for 1fdfb58
  

  3 peopleauthored
  1fdfb58

  Copy full SHA for 1fdfb58

Commits on May 1, 2025

• Add a scattered kv cache. (#2936)

  Show description for cd96fa8
  LaurentMazare

  authored
  cd96fa8

  Copy full SHA for cd96fa8

• Bump the candle version to 0.9.1. (#2935)

  LaurentMazare

  authored
  8a19bb7

  Copy full SHA for 8a19bb7

Commits on Apr 30, 2025

• Add support for Helium-v1. (#2932)

  LaurentMazare

  authored
  38fc866

  Copy full SHA for 38fc866

Commits on Apr 29, 2025

• Added tracing page to the candle book. (#2922)

  Show description for 5029ac5
  greenrazer

  authored
  5029ac5

  Copy full SHA for 5029ac5

Commits on Apr 28, 2025

• Switch Tensor::full to return a contiguous tensor. (#2929)

  LaurentMazare

  authored
  de23d34

  Copy full SHA for de23d34

• Fix the gumbel softmax by casting to f32. (#2928)

  LaurentMazare

  authored
  d4bac37

  Copy full SHA for d4bac37

• Optimize Tensor::new when called on nested Vec<..>. (#2927)

  Show description for e98754f
  LaurentMazare

  authored
  e98754f

  Copy full SHA for e98754f

Commits on Apr 27, 2025

• Support for "unbatched" rope. (#2926)

  Show description for e3db300
  LaurentMazare

  authored
  e3db300

  Copy full SHA for e3db300

• Remove redundant mlx gemm dtype check (#2925)

  ivarflakstad

  authored
  6e0646c

  Copy full SHA for 6e0646c

Commits on Apr 26, 2025

• Bump the crate version to 0.9.0. (#2924)

  LaurentMazare

  authored
  fbaf0b0

  Copy full SHA for fbaf0b0

Pagination

Previous

Next