Commit Graph

  • c2764d1073 Bump actions/download-artifact from 6 to 7 (#2912) main dependabot[bot] 2025-12-15 06:10:16 -08:00
  • 093a62d2ed Bump actions/upload-artifact from 5 to 6 (#2911) dependabot[bot] 2025-12-15 06:09:55 -08:00
  • 1b591ec736 No VJP for mask or sinks in attention (#2909) Awni Hannun 2025-12-13 19:48:39 -08:00
  • 47d2505ea9 Fix attention for large sizes (#2903) Awni Hannun 2025-12-13 06:54:30 -08:00
  • bedefed784 Fix ccache getting disabled (#2905) Cheng 2025-12-13 13:00:51 +09:00
  • d2bc340df4 Update the mlx.launch and mlx.distributed_config docs ibv-backend Angelos Katharopoulos 2025-12-12 17:03:38 -08:00
  • fabc947df4 Possible doc fix Angelos Katharopoulos 2025-12-12 15:34:59 -08:00
  • 5523087cfb Finish the distributed docs Angelos Katharopoulos 2025-12-12 14:16:16 -08:00
  • 2f939acefa Progress with the docs Angelos Katharopoulos 2025-12-12 04:36:46 -08:00
  • ccaaa7d6df fix: possible heap-buffer-overflow in RandomBits::eval_cpu (#2877) Melissa Kilby 2025-12-12 02:11:18 -08:00
  • 3b416a2e36 Comments Angelos Katharopoulos 2025-12-12 01:21:43 -08:00
  • f3e5ca5414 [CUDA] Add host nodes to subgraph types for graph update (#2901) Awni Hannun 2025-12-11 19:13:44 -08:00
  • 753c6a4d0f Disable echo for interactive distributed Angelos Katharopoulos 2025-12-11 17:47:42 -08:00
  • 81dfe5f137 Fix grad in place updates (#2899) Awni Hannun 2025-12-11 14:44:58 -08:00
  • 012fb220a1 fp quantize (#2892) Anastasiia Filippova 2025-12-11 15:11:25 +01:00
  • e1fee0074b Update nanobind pin to most recent version (#2896) Nathan Goldbaum 2025-12-11 07:07:36 -07:00
  • 3c8ce9b00e Fix input buffer donation in compile (#2897) CCYeh 2025-12-11 15:07:03 +01:00
  • 937ce79660 do not use simd neon intrinsics on x86 (#2893) David Koski 2025-12-10 12:23:28 -08:00
  • d3a754c8aa Fix config when more cables are connected Angelos Katharopoulos 2025-12-10 02:14:29 -08:00
  • 595a4ad206 Improve interactivity of mlx.launch Angelos Katharopoulos 2025-12-10 00:54:27 -08:00
  • 208f5441a7 bump minimum required Python version (#2891) Nathan Goldbaum 2025-12-09 17:54:38 -07:00
  • a4dc1fac6c Set the shell to bash explicitly Angelos Katharopoulos 2025-12-09 15:17:09 -08:00
  • ebda161a86 Remove old joined script Angelos Katharopoulos 2025-12-09 13:39:57 -08:00
  • fa31a4b295 Add more checks and improve errors Angelos Katharopoulos 2025-12-09 13:36:17 -08:00
  • 9d707ba3b5 Remove python from the launch script Angelos Katharopoulos 2025-12-09 13:04:37 -08:00
  • b862d842e1 Allow events in sub graph to be updatable (#2886) Awni Hannun 2025-12-09 12:34:37 -08:00
  • f7a400951a Fix docs: replace mx.random.randn with mx.random.normal (#2890) Satyam singh 2025-12-10 01:16:30 +05:30
  • 405d30b6e5 Refactor distributed config Angelos Katharopoulos 2025-12-09 05:58:44 -08:00
  • cd4b12ce1b Refactoring launcher Angelos Katharopoulos 2025-12-01 16:31:57 -08:00
  • 425043ccca Change the name to a fun pun Angelos Katharopoulos 2025-11-20 17:48:23 -08:00
  • 95d92af8a0 Add headers for gcc Angelos Katharopoulos 2025-11-20 17:24:22 -08:00
  • bfdddd644b Expose per-backend availability in C++ and python Angelos Katharopoulos 2025-11-20 15:26:59 -08:00
  • 1216afdc91 Add a no_ibv Angelos Katharopoulos 2025-11-20 12:35:37 -08:00
  • 04e94d78bb Add empty sum_scatter Angelos Katharopoulos 2025-11-20 12:15:27 -08:00
  • 60d4e8b2a8 Add send/recv Angelos Katharopoulos 2025-10-29 14:09:25 -07:00
  • c5745fddd2 Make sure that there is space for work completions Angelos Katharopoulos 2025-10-28 11:27:20 -07:00
  • e937a8033f Add working reduce and semi-working all gather Angelos Katharopoulos 2025-10-17 19:03:26 +03:00
  • 4dfe02d7c6 Fix ring Angelos Katharopoulos 2025-10-15 00:32:22 -07:00
  • 5c2cff9329 Fix side channel initialization for more than 2 peers Angelos Katharopoulos 2025-10-14 17:48:53 -07:00
  • 325dab9559 All gather Angelos Katharopoulos 2025-10-01 01:36:59 -07:00
  • 67e454ab0a Initial working all reduce Angelos Katharopoulos 2025-09-09 13:32:06 -07:00
  • 27232db1ba [CUDA] Enable more graphs to be updatable (#2883) Awni Hannun 2025-12-08 06:18:01 -08:00
  • dd91ee9534 Refactoring launcher ibv-backend-test Angelos Katharopoulos 2025-12-01 16:31:57 -08:00
  • a4b3bc969b Try not to fail when there should be memory available (#2869) Awni Hannun 2025-12-07 06:11:00 -08:00
  • 667c0f3bb9 [Metal] No copy array init (#2875) Awni Hannun 2025-12-05 13:36:45 -08:00
  • 6245824d42 Make allocator::malloc throw on allocation failure (#2874) Cheng 2025-12-05 17:44:38 +09:00
  • 39289ef025 [CUDA] Release build for cuda 13 (#2872) Awni Hannun 2025-12-04 21:42:26 -08:00
  • aefc9bd3f6 [CUDA] Faster general copy (#2873) Awni Hannun 2025-12-04 21:42:15 -08:00
  • 997cfc7699 Add a 2-pass col reduce for CUDA (#2863) Angelos Katharopoulos 2025-12-04 15:53:59 -08:00
  • 1fa8dc5797 Do a PyPi release for cuda on arm (#2866) Awni Hannun 2025-12-04 15:28:29 -08:00
  • 8fab4f0929 Change the name to a fun pun Angelos Katharopoulos 2025-11-20 17:48:23 -08:00
  • 47af2c8cb0 Add headers for gcc Angelos Katharopoulos 2025-11-20 17:24:22 -08:00
  • f40152ebc1 Expose per-backend availability in C++ and python Angelos Katharopoulos 2025-11-20 15:26:59 -08:00
  • 5d7e6a0642 Add a no_ibv Angelos Katharopoulos 2025-11-20 12:35:37 -08:00
  • b9b78b1059 Add empty sum_scatter Angelos Katharopoulos 2025-11-20 12:15:27 -08:00
  • 45727b0c02 Add send/recv Angelos Katharopoulos 2025-10-29 14:09:25 -07:00
  • 2444fbdfe9 Make sure that there is space for work completions Angelos Katharopoulos 2025-10-28 11:27:20 -07:00
  • f3b605e53c Add working reduce and semi-working all gather Angelos Katharopoulos 2025-10-17 19:03:26 +03:00
  • 0388ae3aaf Fix ring Angelos Katharopoulos 2025-10-15 00:32:22 -07:00
  • d4c1de4a8b Fix side channel initialization for more than 2 peers Angelos Katharopoulos 2025-10-14 17:48:53 -07:00
  • 4dbffb3954 All gather Angelos Katharopoulos 2025-10-01 01:36:59 -07:00
  • b1a60b2d2d Initial working all reduce Angelos Katharopoulos 2025-09-09 13:32:06 -07:00
  • a6d6717181 fix compile copying (#2871) Awni Hannun 2025-12-04 12:32:56 -08:00
  • 941cfe23d7 Layer norm throws on dimension mismatch (#2870) Awni Hannun 2025-12-04 11:21:05 -08:00
  • 9abb0b8123 Added support for pytree types that inherit from tuple and typing.namedtuple (#2845) romanoneg 2025-12-04 11:06:45 -08:00
  • 50d3914c67 Update gumbel function signature parameters (#2868) Tian En "TianHeng 2025-12-03 23:37:35 +00:00
  • cacbdbf995 Fix init from double (#2861) Awni Hannun 2025-12-03 06:08:11 -08:00
  • 193cdcd81a Fix graph updating (#2857) Awni Hannun 2025-12-02 17:12:24 -08:00
  • d8ceae7b77 Reduce JVP (#2854) Awni Hannun 2025-12-02 16:17:47 -08:00
  • 5cf6f10bef Add debug line info jit-nax Jagrit Digani 2025-12-02 14:49:11 -08:00
  • 7c1abc50c0 Update make compiled preamble to not preprocess macros Jagrit Digani 2025-12-02 14:25:00 -08:00
  • eff0e31f00 Fix export scatters (#2852) Awni Hannun 2025-12-02 11:24:40 -08:00
  • 6c5785bc2f use thread local cpature mode (#2850) Awni Hannun 2025-12-01 19:02:47 -08:00
  • 8879ee00eb Support more Numpy interfaces for masked_scatter (#2832) CCYeh 2025-12-02 02:51:02 +01:00
  • 6e762fe2e2 [CUDA] Migrate conv code to new cuDNN APIs (#2847) Cheng 2025-12-02 07:55:43 +09:00
  • 2b95d0c270 [CUDA] Use cuDNN attention when T_q != T_kv (#2843) Cheng 2025-11-27 09:58:43 +09:00
  • b054838780 Added clarification to apply_fn parameter of apply_to_modules (#2831) Chaoran Yu 2025-11-26 15:40:56 -08:00
  • dd79d3c465 [CUDA] Faster rms norm for small dimension (#2838) Awni Hannun 2025-11-26 15:10:41 -08:00
  • 704fd1ae28 [CUDA] Support array mask in SDPA (#2822) Cheng 2025-11-26 11:08:58 +09:00
  • c9f4dc851f Merge build-cuda and build-linux actions (#2783) Cheng 2025-11-25 20:06:42 +09:00
  • f8bd675655 [CUDA] Output of SDPA should have same layout with inputs (#2826) Cheng 2025-11-25 15:22:58 +09:00
  • 23a9168d34 [CUDA] Add debug env to save cuda graphs to dot files (#2825) Cheng 2025-11-25 15:22:36 +09:00
  • bca205e287 [CUDA] Exit on crash and more helpful errors (#2830) Awni Hannun 2025-11-24 19:46:03 -08:00
  • 1d4eacb737 Fix mx.core.linspace type annotation (#2820) CCYeh 2025-11-24 23:15:08 +01:00
  • 8abd37ad05 Bump actions/checkout from 5 to 6 (#2828) dependabot[bot] 2025-11-24 06:04:46 -08:00
  • 3e05cea9f8 Force cudaGraphExec reinstantiation when clusters are used (#2813) Andrey Portnoy 2025-11-22 15:43:49 -05:00
  • 5b0f047226 Fix mx.core.load type annotation (#2819) CCYeh 2025-11-22 20:09:44 +01:00
  • 618c87af8c Add float64 Eig and complex64 SVD/Eig support (Fixes #2708) (#2737) Harsh Sutaria 2025-11-22 09:51:36 -05:00
  • d5f61a93fa Fix typo: refs/head/main => refs/heads/main (#2818) Cheng 2025-11-22 09:43:35 +09:00
  • 4a09264236 Tolerance for some ops tests on cuda (#2815) Awni Hannun 2025-11-21 16:06:16 -08:00
  • 0dbc7e5bee Centralize NAX condition (#2811) Awni Hannun 2025-11-21 13:28:15 -08:00
  • 0d68efd461 patch bump for future version (#2804) Awni Hannun 2025-11-20 09:26:20 -08:00
  • f9e1a14135 [CUDA] Partly fix random for large sizes (#2798) Awni Hannun 2025-11-20 07:27:50 -08:00
  • d8e9ded928 Fix cuda allocator copy condition (#2800) Awni Hannun 2025-11-20 07:06:55 -08:00
  • 60939d010c Fix macos release target and linux arm release (#2802) Awni Hannun 2025-11-19 21:37:50 -08:00
  • fdcd2923fd patch + fix docs build (#2799) Awni Hannun 2025-11-19 16:16:26 -08:00
  • 54f1cc6e3e Add Neural Accelerator Support (#2772) v0.30.0 Jagrit Digani 2025-11-19 15:06:00 -08:00
  • b3825ac149 Add Masked Scatter (#2663) CCYeh 2025-11-19 23:53:32 +01:00
  • 7f4b7e553c version (#2797) Awni Hannun 2025-11-19 14:11:16 -08:00
  • ad16f41a7f Fix version tag (#2790) Awni Hannun 2025-11-19 08:55:57 -08:00