Commit Graph

  • 3d7b81535a
    Only link cuda driver librairies. Nicolas Patry 2024-09-13 16:50:29 +0200
  • e898483db6
    Updating outlines to 0.0.46 Nicolas Patry 2024-09-13 15:38:38 +0200
  • ce3efc83ed
    Remove tmate. Nicolas Patry 2024-09-12 16:28:01 +0200
  • 7f58f7dc61
    Symlink all the things. Nicolas Patry 2024-09-12 16:23:33 +0200
  • 42107de71f
    Let's try to find libnvidia-ml Nicolas Patry 2024-09-12 16:16:29 +0200
  • edaa7f847d
    Does this work ? Nicolas Patry 2024-09-12 16:14:45 +0200
  • d1e79ddae0
    Fix override. Nicolas Patry 2024-09-12 16:08:04 +0200
  • db054b95df
    Check the paths. Nicolas Patry 2024-09-12 15:49:05 +0200
  • afcd047a58
    Yaml yaml. Nicolas Patry 2024-09-12 15:44:12 +0200
  • 60db294f9a
    Link cuda to nix ? Nicolas Patry 2024-09-12 15:38:57 +0200
  • 8e7c7c61f1
    Let's see what the issue is ? Nicolas Patry 2024-09-12 15:21:05 +0200
  • 815449da74
    Removing unused code. Nicolas Patry 2024-09-12 15:07:28 +0200
  • c227345878
    Run on actual GPUs. Nicolas Patry 2024-09-12 14:52:36 +0200
  • 3d73c99ebe
    Attempt at integration tests. Nicolas Patry 2024-09-12 14:47:18 +0200
  • f47cdc1fe1
    Attempting rapidly the integration tests. Nicolas Patry 2024-09-12 14:27:48 +0200
  • 38fcafcf96
    Adding a test for FD. (#2516) Nicolas Patry 2024-09-16 17:00:54 +0200
  • 5726a9ca81 Move to moe-kernels package and switch to common MoE layer Daniël de Kok 2024-09-16 09:45:48 +0000
  • 7774655297
    Add tests for Mixtral (#2520) Daniël de Kok 2024-09-16 12:39:18 +0200
  • 5f82151e2e Add tests for Mixtral Daniël de Kok 2024-09-13 13:01:34 +0000
  • 07efcf423e
    Increasing docker timeout. Nicolas Patry 2024-09-14 18:25:15 +0200
  • 7d7fa19147
    Update the locks. Nicolas Patry 2024-09-14 16:29:58 +0200
  • 6769e45711
    Update hash for slice.len() == 1 Nicolas Patry 2024-09-14 15:06:03 +0200
  • 5fa332156f
    Use an actual hash. Nicolas Patry 2024-09-13 18:34:53 +0200
  • 9e45a09a0a
    Last reference. Nicolas Patry 2024-09-13 18:23:06 +0200
  • b043f56ed2
    Fixing radix with block_size > 1 Nicolas Patry 2024-09-13 18:01:56 +0200
  • 4fc01e243d
    Fixing the invalid popping. Nicolas Patry 2024-09-13 16:49:28 +0200
  • a08f7eb993
    Fixing flashdecoding (empty batch doesn't work). Nicolas Patry 2024-09-12 17:26:53 +0200
  • f6697baf31
    Adding a test for FD. Nicolas Patry 2024-09-12 11:12:18 +0200
  • 9cca3e0b03
    Use ratatui not (deprecated) tui (#2521) Alex Strick van Linschoten 2024-09-13 18:45:28 +0200
  • 10628e878a Merge branch 'main' into gpt_awq_4 pr-2444-ci-branch Wang, Yi A 2024-09-13 04:45:19 -0400
  • 4ba9210f91 fix docker Mohit Sharma 2024-09-12 15:45:06 +0000
  • 3ac7df2b6d
    hotfix : enable intel ipex cpu and xpu in python3.11 (#2517) Wang, Yi 2024-09-12 23:23:49 +0800
  • 628334d336
    fix: pass missing revision arg for lora adapter when loading multiple… (#2510) drbh 2024-09-12 17:04:52 +0200
  • 59fd0cbdff add skinny kernel and merge fixes Mohit Sharma 2024-09-12 13:16:13 +0000
  • d95c670ada
    Add nix test. (#2513) Nicolas Patry 2024-09-12 14:54:56 +0200
  • 6b995cca30 enable intel ipex cpu and xpu in python3.11 pr-2517-ci-branch Wang, Yi A 2024-09-12 05:47:26 -0700
  • a95084d5ea
    Ignore the cache for now. Nicolas Patry 2024-09-12 14:19:22 +0200
  • 804715216e
    Attempting to use a cache location for the models. Nicolas Patry 2024-09-12 13:59:53 +0200
  • 16f71106c2
    Up. Nicolas Patry 2024-09-12 13:52:59 +0200
  • e60bdb2c89
    Update it a bit. Nicolas Patry 2024-09-12 13:51:11 +0200
  • eaf2533d1a
    Test requires cargo for cargo fmt. Nicolas Patry 2024-09-12 13:45:35 +0200
  • 89583ed0f3
    Missing pre-commit. Nicolas Patry 2024-09-12 13:41:58 +0200
  • 0621dfce80
    Adding the other tests. Nicolas Patry 2024-09-12 13:31:48 +0200
  • 533d9991e4
    Fixed the auth token ? Nicolas Patry 2024-09-12 12:24:36 +0200
  • b7f1129f92
    Add the secrets. Nicolas Patry 2024-09-12 11:55:42 +0200
  • 21832a1d4d
    Add a formatter. Nicolas Patry 2024-09-12 11:44:10 +0200
  • ca934411d1
    Forgot this modification. Nicolas Patry 2024-09-12 11:36:43 +0200
  • 46efc844d6
    Add the actual test target. Nicolas Patry 2024-09-12 11:35:48 +0200
  • 75486de71d
    Different user ? Nicolas Patry 2024-09-12 11:27:22 +0200
  • 21440a47a3
    Root user. Nicolas Patry 2024-09-12 11:23:34 +0200
  • 952f503332
    Reemove server. Nicolas Patry 2024-09-12 11:16:37 +0200
  • 5bd547cdf1
    Our runner + pure test (not written) Nicolas Patry 2024-09-12 11:15:47 +0200
  • ca7267da0e
    Try thuis. Nicolas Patry 2024-09-11 18:11:29 +0200
  • ead590b2af
    Fixing the test + adding click (needed for pre-commit hooks). Nicolas Patry 2024-09-11 18:00:54 +0200
  • bffac7396a
    Modifying yourself means you need to rerun. Nicolas Patry 2024-09-11 17:27:09 +0200
  • 4b6c723f6d
    Add nix test. Nicolas Patry 2024-09-11 17:22:01 +0200
  • 94304649f1
    nix: support Python tokenizer conversion in the router (#2515) Daniël de Kok 2024-09-12 10:44:01 +0200
  • 905e4b9ac8 nix: support Python tokenizer conversion in the router Daniël de Kok 2024-09-12 08:34:37 +0000
  • 69e3be20fb
    Fix truffle (#2514) Nicolas Patry 2024-09-11 22:45:19 +0200
  • dae3bf1d87
    Fix tokenization yi (#2507) Nicolas Patry 2024-09-11 22:41:56 +0200
  • 555bdaa72e
    Attempt to fix trufflehog. Nicolas Patry 2024-09-11 22:13:01 +0200
  • b181ee37f3
    Attempting to discard the trufflehog warning. Nicolas Patry 2024-09-11 21:46:54 +0200
  • 730ccb9090
    Forcing 3.11 ? Nicolas Patry 2024-09-11 21:08:34 +0200
  • bc4b0c2c70
    Why do we want mkl on AMD ? Nicolas Patry 2024-09-11 18:27:23 +0200
  • 725f86716b
    Put back rust tests. Nicolas Patry 2024-09-11 18:18:18 +0200
  • 538994ced4
    Updating the dockerfile to make libpython discoverable at runtime too. Nicolas Patry 2024-09-11 16:45:22 +0200
  • f518f798b7
    Apparently 3.10 is not available anymore. Nicolas Patry 2024-09-11 16:02:30 +0200
  • eb4d6b06e2
    -y. Nicolas Patry 2024-09-11 15:19:55 +0200
  • c1c207206d
    WTF. Nicolas Patry 2024-09-11 15:18:05 +0200
  • d31fff8ac3
    Desperation. Nicolas Patry 2024-09-11 14:59:40 +0200
  • 38e9349493
    Tmate the hell out of this. Nicolas Patry 2024-09-11 14:36:50 +0200
  • ae88fa2f61
    Shot in the dark. Nicolas Patry 2024-09-11 14:32:58 +0200
  • d40b4ea675
    Tmp. Nicolas Patry 2024-09-11 13:16:55 +0200
  • 58f3b66556
    have no idea at this point Nicolas Patry 2024-09-11 13:14:09 +0200
  • 5cf4336af6
    Monkey it up. Nicolas Patry 2024-09-11 13:07:28 +0200
  • e04f480ad2
    List stuff. Nicolas Patry 2024-09-11 12:55:46 +0200
  • 5298fd0116
    Getting libpython maybe ? Nicolas Patry 2024-09-11 12:52:14 +0200
  • e158141a83
    No sccache. Nicolas Patry 2024-09-11 12:47:48 +0200
  • 1d3d655ac8
    Remove sccache Nicolas Patry 2024-09-11 12:43:41 +0200
  • 0699faf46e
    Upgrade python version. Nicolas Patry 2024-09-11 12:42:04 +0200
  • 8d473a568e
    Try a faster runner Nicolas Patry 2024-09-11 12:40:28 +0200
  • 1c296d6047
    Validation is odd. Nicolas Patry 2024-09-11 12:12:12 +0200
  • 33683220ca
    Fixing the location ? Nicolas Patry 2024-09-11 12:10:24 +0200
  • 50c8ca6c84
    Fix the gh action? Nicolas Patry 2024-09-11 11:54:12 +0200
  • e78723330e
    Fixing the builds ? Nicolas Patry 2024-09-11 11:51:19 +0200
  • b3760df25f
    Fixing odd tokenization self modifications on the Rust side (load and resave in Python). Nicolas Patry 2024-09-10 19:11:32 +0200
  • a4e3e8c608
    Prefix test - Different kind of load test to trigger prefix test bugs. (#2490) Nicolas Patry 2024-09-11 18:10:40 +0200
  • d930f11621
    We want only 1 run of integration tests..... Nicolas Patry 2024-09-11 16:46:05 +0200
  • 397a10aa22 fix: pass missing revision arg for lora adapter when loading multiple adapters drbh 2024-09-11 13:44:52 +0000
  • 5c0a252e63
    Fingers crossed. Nicolas Patry 2024-09-11 15:34:59 +0200
  • 058162685f fixed merge conflicts Mohit Sharma 2024-09-11 10:57:40 +0000
  • 0345816477 fix mixtral model Mohit Sharma 2024-09-11 10:52:10 +0000
  • 1c3ef1c184
    Tmate again. Nicolas Patry 2024-09-11 11:58:02 +0200
  • e2f48fae3d hide env vart Mohit Sharma 2024-09-11 07:00:29 +0000
  • f3bc038430 style Mohit Sharma 2024-09-11 06:52:30 +0000
  • 7b13fede50
    Important line got squashed. Nicolas Patry 2024-09-10 18:26:16 +0200
  • aac91d1d26
    Updating VLM causal model with updated context. Nicolas Patry 2024-09-10 18:00:21 +0200
  • 1bae63973a
    Praying. Nicolas Patry 2024-09-10 17:09:22 +0200
  • 50a58fe9dc
    Tmate. Nicolas Patry 2024-09-10 16:45:45 +0200
  • c85eeb18cc
    wip Guillaume LEGENDRE 2024-09-10 16:23:49 +0200