Commit Graph

  • 90b9f450c4
    Merge 0c019b21f4 into e51eacecb0 #2565 Jack Hogan 2025-02-20 10:12:14 +0000
  • a745ea3772
    Merge 7291b0a8e9 into e51eacecb0 #2623 James Hiew 2025-02-20 15:29:57 +0800
  • ca23fb50b4
    Merge 9507624c65 into e51eacecb0 #2830 ZC 2025-02-20 02:46:36 +0000
  • 9507624c65 burn-autodiff: support no_std #2830 ivila 2025-02-19 12:13:50 +0800
  • f8e9a168b4
    Merge 181c3d5b18 into e51eacecb0 #2828 Xwg 2025-02-20 00:50:40 +0000
  • 181c3d5b18 Remove unused import `TensorType` from gather.rs #2828 Xwg 2025-02-20 08:50:22 +0800
  • bf2ca68a61
    Merge 76a4a88aa7 into e51eacecb0 #2786 Anthony Torlucci 2025-02-19 15:12:59 -0600
  • 0c8d4fb608
    Merge 08db3065a6 into e51eacecb0 #2831 Dilshod Tadjibaev 2025-02-19 14:44:58 -0600
  • 08db3065a6 Rename dim to rank #2831 Dilshod Tadjibaev 2025-02-19 14:42:06 -0600
  • 1b29302400 Fix swap dims feat/vectorization-fusion nathaniel 2025-02-19 13:34:14 -0500
  • d101d0c7ea add feature flag for wasm remove-wasm-shared-sum maxtremblay 2025-02-19 10:56:41 -0500
  • b89d674b54 burn-core: support no-std optim ivila 2025-02-19 12:09:06 +0800
  • 322fb210d2 Refactor GatherNode to support scalar outputs. Xwg 2025-02-19 10:39:32 +0800
  • 5a542374bf It works, except tensor views nathaniel 2025-02-18 17:13:17 -0500
  • 3c16252176 WIP nathaniel 2025-02-18 15:32:37 -0500
  • 30eefdfc3d Add per-block qparams compute feat/quant/per-block Guillaume Lagrange 2025-02-18 13:47:56 -0500
  • 8f7ca98584 Add per-block calibration range Guillaume Lagrange 2025-02-18 12:11:41 -0500
  • 6ea784aa14 Refactor scheme to have quantization mode + add per-block variant Guillaume Lagrange 2025-02-14 15:59:33 -0500
  • e51eacecb0
    Proposal to expose ATOL/RTOL defaults (#2824) main Crutcher Dunnavant 2025-02-18 11:18:09 -0800
  • 7de2a5cf33 flip #2824 Crutcher Dunnavant 2025-02-18 10:42:41 -0800
  • 2c84b8c6f5 Review Crutcher Dunnavant 2025-02-18 10:40:54 -0800
  • 3810766979 Merge branch 'main' into feat/vectorization-fusion nathaniel 2025-02-18 13:33:42 -0500
  • 96839b99f6
    Cleanup Tensor Registry in fusion (#2826) Nathaniel Simard 2025-02-18 13:15:05 -0500
  • 32f3cdfc36 Clippy #2826 nathaniel 2025-02-18 12:14:04 -0500
  • 46792b2499 WIP nathaniel 2025-02-18 12:08:55 -0500
  • 0798b9173b Cleanup Tensor Registry in fusion nathaniel 2025-02-18 11:22:21 -0500
  • 610943da25
    Add data_dir optional argument to Huggingface DataLoader to enable some manual download use cases (#2817) Paweł Biłko 2025-02-18 06:22:45 -0800
  • 1604a61f11 Proposal to expose ATOL/RTOL defaults Crutcher Dunnavant 2025-02-17 15:38:41 -0800
  • f3dfd05c6a
    Refactor/cubecl fusion (#2815) Nathaniel Simard 2025-02-17 17:38:59 -0500
  • bbcb438728 Cleanup #2815 nathaniel 2025-02-17 17:13:42 -0500
  • 4e64fb0a5a Fix burn-core backend testing dependencies nathaniel 2025-02-17 16:45:51 -0500
  • 0a1e515fc6 Rename matmul;2u nathaniel 2025-02-17 16:36:34 -0500
  • a7e6042e0a Add publish github + cargo description nathaniel 2025-02-17 16:32:08 -0500
  • 1505fe5c31 Clippy + Fmt nathaniel 2025-02-17 16:24:48 -0500
  • ef29e1f453 Fix conv nathaniel 2025-02-17 16:18:29 -0500
  • 5b5a188c36 Merge branch 'main' into refactor/cubecl-fusion nathaniel 2025-02-17 16:14:38 -0500
  • 697a4889a5 Merge branch 'main' into refactor/cubecl-fusion nathaniel 2025-02-17 16:11:26 -0500
  • 136eeb6fb7
    bump cubecl version with dummy implementations (#2814) Maxime Tremblay 2025-02-17 16:11:11 -0500
  • b4bc1b9779 Update rev nathaniel 2025-02-17 16:06:17 -0500
  • c1bb1abdd4 Fix scalar indexing nathaniel 2025-02-17 15:48:11 -0500
  • 46008ac18a Fix a few things nathaniel 2025-02-17 15:07:13 -0500
  • 9ae90c2af7 It compiles nathaniel 2025-02-17 13:33:14 -0500
  • 37822fdb51
    Feat/Split ONNX Import (#2568) Mathias Insley 2025-02-17 07:28:36 -0800
  • 163a2069c6
    Optimize connected components (#2810) Genna Wingert 2025-02-17 15:51:38 +0100
  • a6cd17d05e
    Bump tempfile to 3.16.0 (#2821) Guillaume Lagrange 2025-02-17 09:49:11 -0500
  • 542c06c1f2 Bump tempfile to 3.16.0 #2821 Guillaume Lagrange 2025-02-17 09:24:07 -0500
  • 56146514e2
    Combined PRs (#2820) github-actions[bot] 2025-02-17 08:56:33 -0500
  • de33698a0c
    Merge dependabot/cargo/candle-core-0.8.3 into combined-prs-branch #2820 github-actions[bot] 2025-02-17 06:06:59 +0000
  • c457634cc2
    Merge dependabot/cargo/tempfile-3.17.0 into combined-prs-branch github-actions[bot] 2025-02-17 06:06:59 +0000
  • 11ca40c4d5
    Bump tempfile from 3.16.0 to 3.17.0 #2819 dependabot[bot] 2025-02-17 02:57:43 +0000
  • 3e31f404f8
    Bump candle-core from 0.8.2 to 0.8.3 #2818 dependabot[bot] 2025-02-17 02:55:16 +0000
  • a823fe8474 Add data_dir optional argument to Huggingface DataLoader to enable some manual download use cases #2817 Pawel Bilko 2025-02-16 15:20:24 -0800
  • 5e0c3213ce Move code for generating uninitialized tensors into burn-import #2568 Mathias 2025-02-14 14:32:02 -0800
  • 1e2a013182 Remove initializer helpers Mathias 2025-02-14 14:31:18 -0800
  • 286ff26a4b Use polyfill in fusion nathaniel 2025-02-14 17:21:23 -0500
  • 15143d9574 Move backends to burn root crate nathaniel 2025-02-14 16:28:35 -0500
  • 59451ace5d Cleanup nathaniel 2025-02-14 15:55:57 -0500
  • d9f479e71b Update readmes nathaniel 2025-02-14 15:51:37 -0500
  • 4bf67b1972 Cleanup nathaniel 2025-02-14 15:48:18 -0500
  • 5ca92e35fd It compiles nathaniel 2025-02-14 14:58:33 -0500
  • c74f658aa4 bump cubecl version with dummy implementations #2814 maxtremblay 2025-02-14 14:13:44 -0500
  • 5052bc69d2 Fix hardcoded `DType::Bool` uses #2810 Genna Wingert 2025-02-14 19:37:30 +0100
  • 3f979cbd72 WIP nathaniel 2025-02-14 13:34:04 -0500
  • 76a4a88aa7 fix clippy lint suggestions #2786 anthonytorlucci 2025-02-14 12:19:55 -0600
  • 51446106d6 Fix display for bool Genna Wingert 2025-02-14 18:49:42 +0100
  • a3aaea4611 Add `to_bool` and fix `Bool::Elem` Genna Wingert 2025-02-14 18:41:01 +0100
  • 7b564f8a3a updated unet example with suggested changes anthonytorlucci 2025-02-14 10:45:10 -0600
  • 9abe891373 Fix and/or tests Genna Wingert 2025-02-14 16:44:27 +0100
  • 7b83129e9c Merge branch 'main' into opt/connected-components Genna Wingert 2025-02-14 16:43:57 +0100
  • 2c4c03941f
    Add boolean and/or to bool tensors (#2802) Genna Wingert 2025-02-14 15:53:35 +0100
  • f72a656a6e Fix nn/mask tests Genna Wingert 2025-02-14 15:38:21 +0100
  • 38aeb34ad8 Don't assert type equality for bool tests Genna Wingert 2025-02-14 15:18:38 +0100
  • d1453f1cf5
    Update crates/burn-fusion/src/stream/context.rs #2802 Genna Wingert 2025-02-14 14:37:12 +0100
  • 45d9e43c99 Fill in placeholder docstrings Mathias 2025-02-14 00:17:35 -0800
  • 3730f52902 Copy override logic from @jameshiew Mathias 2025-02-14 00:10:20 -0800
  • 433ead8570 Revert attempt at sidestepping constant int node issue Mathias 2025-02-14 00:08:34 -0800
  • b4a64ffe44 Intraconvert from ONNX graph node to SplitConfig properly Mathias 2025-02-14 00:07:45 -0800
  • 0cb6903de4 Change num_outputs to split_size in SplitConfig to follow burn implementation Mathias 2025-02-14 00:07:10 -0800
  • 00f50c40a2 Potential workaround to constant int node Mathias 2025-02-13 16:53:06 -0800
  • 6a91f0cadd Update onnx test now that return is tuple of static size Mathias 2025-02-13 15:18:35 -0800
  • 2d01f4a448 Don't convert to bool in place Genna Wingert 2025-02-14 00:10:25 +0100
  • 8e0e74b476 `bool_into_data` should have a dtype of `BoolElem` Genna Wingert 2025-02-13 23:56:19 +0100
  • 07f64e8299 Cleanup build errors Mathias 2025-02-13 14:48:33 -0800
  • f9209a0d45 Make lint `allow` because it's buggy Genna Wingert 2025-02-13 23:44:34 +0100
  • 588225624b Merge branch 'main' into feat/bool-and-or Genna Wingert 2025-02-13 23:27:37 +0100
  • edab72011b Merge branch 'main' into opt/connected-components Genna Wingert 2025-02-13 23:22:22 +0100
  • b406082efb Use existing into_data implementation Genna Wingert 2025-02-13 23:18:26 +0100
  • 13e75e91ce Optimize NdArray into_data and other ways to make connected components faster Genna Wingert 2025-02-13 23:10:00 +0100
  • 6f52bc84f6 Update supported ops list Mathias 2025-02-13 13:02:43 -0800
  • 45c316b573 Clippy fixes Mathias 2025-02-13 13:01:00 -0800
  • be0cce89dc Keep the number of output tensor sizes fixed Mathias 2025-02-13 12:58:47 -0800
  • d9e41460ff
    Refactor burn jit => burn-cubecl (#2809) Nathaniel Simard 2025-02-13 12:39:29 -0500
  • 5c35d44a68 More rename #2809 nathaniel 2025-02-13 12:11:51 -0500
  • 74dcfe77d3 Update workflow nathaniel 2025-02-13 12:05:37 -0500
  • f53de57868 Merge branch 'main' into refactor/burn-jit nathaniel 2025-02-13 12:04:41 -0500
  • 3ed38cfcb9 Rename crate nathaniel 2025-02-13 11:59:09 -0500
  • 18796a7c83 Cleanup nathaniel 2025-02-13 11:46:30 -0500
  • d91929d063 Rename JitAutotuneKey => CubeAutotuneKey nathaniel 2025-02-13 11:42:39 -0500
  • 5a4483b529 Rename JitElement => CubeElement nathaniel 2025-02-13 11:41:23 -0500
  • 5ed227d4bd Rename JitFusionHandle => CubeFusionHandle nathaniel 2025-02-13 11:40:38 -0500