Commit Graph

  • 6afea77ac3 added gloo groups for transfer cpu tensors zbian 2022-03-31 23:05:08 +0800
  • 092b2b015f
    Merge branch 'hpcaitech:main' into feature/monitoring SMesForoush 2022-03-31 18:32:41 +0430
  • 1f31e4f924 add sample.env file Sahel Mesforoush 2022-03-31 18:31:40 +0430
  • 759f021288 fix readme bug Sahel Mesforoush 2022-03-31 18:29:44 +0430
  • cba3fee5b0
    fix format (#586) doubleHU 2022-03-31 21:37:17 +0800
  • 0d6f39e355 fix format huxin711 2022-03-31 21:34:58 +0800
  • db3834cde0
    fix format (#585) fanjinfucool 2022-03-31 20:36:09 +0800
  • aac4e61d6f fix format fanjifu 2022-03-31 19:32:23 +0800
  • 104cbbb313
    [hotfix] add hybrid adam to __init__ (#584) ver217 2022-03-31 19:08:34 +0800
  • 2a31ddafb1 add hybrid adam to __init__ ver217 2022-03-31 19:05:39 +0800
  • e6d50ec107
    [zero] adapt zero for unsharded parameters (#561) HELSON 2022-03-31 18:34:11 +0800
  • 21b638fc2f
    fix format (#583) binmakeswell 2022-03-31 18:10:03 +0800
  • ef8eeb22f1 fix format binmakeswell 2022-03-31 17:58:29 +0800
  • 9955d9c9bb adapt zero for unsharded parameters 1SAA 2022-03-30 00:10:38 +0800
  • 00119f648e polish jiaruifang 2022-03-31 17:45:03 +0800
  • 13ed4b6441
    [model zoo] add activation offload for gpt model (#582) LuGY 2022-03-31 17:42:20 +0800
  • a5eedce109 add activation offload for gpt model lclgy 2022-03-31 17:36:13 +0800
  • 39a0bc920b
    Update test_tensor_move.py FredHuang99 2022-03-31 17:15:50 +0800
  • 46c9ba33da update code format Wesley 2022-03-31 17:11:03 +0800
  • 666cfd094a fix parallel_input flag for Linear1D_Col gather_output Wesley 2022-03-31 16:38:14 +0800
  • 865fc24119
    block_reduce.h fix format (#581) DouJS 2022-03-31 17:13:09 +0800
  • ba1cf5b04a
    fix format (#580) Maruyama_Aya 2022-03-31 17:12:24 +0800
  • 807f5817ae update code format Wesley 2022-03-31 17:11:03 +0800
  • 41a056f149 block_reduce.h fix format dujiangsu 2022-03-31 17:04:14 +0800
  • 68ccda6ae7 polish code jiaruifang 2022-03-31 16:58:16 +0800
  • 46fe597ab0 polish jiaruifang 2022-03-31 16:55:50 +0800
  • f583ac20a7 fix format maruyama 2022-03-31 16:52:21 +0800
  • a2e78b2252 fix parallel_input flag for Linear1D_Col gather_output Wesley 2022-03-31 16:38:14 +0800
  • af40cae46c polish jiaruifang 2022-03-31 16:35:02 +0800
  • a9f778f1b1
    [tool] create .clang-format for pre-commit (#578) BoxiangW 2022-03-31 16:34:00 +0800
  • 253fd1d3f7
    Create .clang-format BoxiangW 2022-03-31 16:32:03 +0800
  • 0e4c861b59 polish jiaruifang 2022-03-31 16:31:59 +0800
  • cf73e684be Merge branch 'main' of https://github.com/hpcaitech/ColossalAI into jiaruifang/memory_utils jiaruifang 2022-03-31 16:28:46 +0800
  • 7c6c427db1
    [zero] trace states of fp16/32 grad and fp32 param (#571) ver217 2022-03-31 16:26:54 +0800
  • 1ac92d02e4 [utils] polish memory utils. jiaruifang 2022-03-31 16:25:45 +0800
  • 0bdb9ae37d
    Update test_tensor_move.py FredHuang99 2022-03-31 16:03:08 +0800
  • 5412186515
    Update test_tensor_move.py FredHuang99 2022-03-31 15:59:22 +0800
  • 20820bde19
    fix format (#574) wky 2022-03-31 15:47:51 +0800
  • 935da1fc66
    fix format (#572) BoxiangW 2022-03-31 15:47:24 +0800
  • 951d7bcaad
    'fix/format' (#573) yuxuan-lou 2022-03-31 15:46:11 +0800
  • 9122ff30ce 'fix/format' yuxuan-lou 2022-03-31 15:44:49 +0800
  • 82133a0ee7 fix format BoxiangW 2022-03-31 15:43:41 +0800
  • 41f26d24c1 fix format wangkuangyi 2022-03-31 15:42:43 +0800
  • 1bf6e23162 polish code ver217 2022-03-31 15:41:54 +0800
  • 15da00ea9c trace states of fp16/32 grad and fp32 param ver217 2022-03-31 15:37:57 +0800
  • 5d533d814f
    fix format (#570) Kai Wang (Victor Kai) 2022-03-31 15:26:41 +0800
  • d6d8020729
    Update test_tensor_move.py FredHuang99 2022-03-31 15:22:10 +0800
  • ed7c704981 fix format kaiwang960112 2022-03-31 15:17:53 +0800
  • a68bef960a
    fix format (#568) Xu Kai 2022-03-31 15:13:01 +0800
  • 238724baf8 [polish] polish MOE gradient handler. jiaruifang 2022-03-31 15:12:33 +0800
  • 714ea7be99
    fix format (#567) YuliangLiu0306 2022-03-31 15:05:58 +0800
  • b95c84c5e0
    [format]colossalai/kernel/cuda_native/csrc/layer_norm_cuda.cpp (#566) Jie Zhu 2022-03-31 15:01:51 +0800
  • 8dc9597545
    fix format (#564) coder-chin 2022-03-31 15:00:50 +0800
  • f040353a23
    fix format (#565) Luxios22 2022-03-31 15:00:21 +0800
  • 2cbccd1c20 [format]colossalai/kernel/cuda_native/csrc/layer_norm_cuda.cpp Jie Zhu 2022-03-31 14:57:41 +0800
  • f66f2d6452 fix format chin 2022-03-31 14:57:18 +0800
  • 47acf3c00e fix format Luxios22 2022-03-31 14:53:00 +0800
  • 4c3a26ea10 fix format liuyuliang 2022-03-31 14:52:51 +0800
  • 04a4edc119
    fix format (#563) Ziyue Jiang 2022-03-31 14:50:16 +0800
  • 376faaf991 fix format Xu Kai 2022-03-31 14:48:17 +0800
  • df78e2423d fix format Wesley 2022-03-31 14:46:21 +0800
  • 09982390ec
    Update test_tensor_move.py FredHuang99 2022-03-31 13:42:09 +0800
  • 7675366fce
    [polish] rename col_attr -> colo_attr (#558) Jiarui Fang 2022-03-31 12:25:45 +0800
  • f7df729e95
    Update test_tensor_move.py FredHuang99 2022-03-31 11:49:06 +0800
  • 2c45efc398
    html refactor (#555) Liang Bowen 2022-03-31 11:36:56 +0800
  • 9ceaea7077 html refactor number1roy 2022-03-30 15:26:09 +0800
  • c26335bc6d [polish] rename col_attr -> colo_attr jiaruifang 2022-03-31 10:37:25 +0800
  • d1211148a7
    [utils] update colo tensor moving APIs (#553) Jiarui Fang 2022-03-30 23:13:24 +0800
  • 80757f7a39 add README for logging-monitoring Sahel Mesforoush 2022-03-30 19:09:52 +0430
  • b836af85fe setup promtail configurations Sahel Mesforoush 2022-03-30 19:08:36 +0430
  • 3f2778b97a add loki-promtail docker compose Sahel Mesforoush 2022-03-30 19:07:54 +0430
  • c44d797072
    [docs] updatad docs of hybrid adam and cpu adam (#552) LuGY 2022-03-30 18:14:59 +0800
  • 014bac0c49
    [zero] hijack p.grad in sharded model (#554) ver217 2022-03-30 18:14:50 +0800
  • 1c90b25e7c polish code jiaruifang 2022-03-30 18:11:07 +0800
  • 8d252bf49d polish comments ver217 2022-03-30 17:53:34 +0800
  • 1669964342 polish comments ver217 2022-03-30 17:43:38 +0800
  • b32f6ac374 hijack p.grad in sharded model ver217 2022-03-30 17:39:46 +0800
  • 16fc2b90ad [utils] update colo tensor moving apis jiaruifang 2022-03-30 17:00:30 +0800
  • f6d6b7cf73 Polish the docs lclgy 2022-03-30 16:14:32 +0800
  • f552b11294
    [zero] label state for param fp16 and grad (#551) Jiarui Fang 2022-03-30 15:57:46 +0800
  • b14321fc1b polish jiaruifang 2022-03-30 15:39:45 +0800
  • 97292c8fb6 [docs] updatad docs of hybrid adam and cpu adam lclgy 2022-03-30 15:27:48 +0800
  • 9d319bf685 polish jiaruifang 2022-03-30 14:47:51 +0800
  • 77443faf03 add pre/post fwd operations jiaruifang 2022-03-30 14:19:00 +0800
  • 1a4d05805b set param fp16 state. jiaruifang 2022-03-30 14:15:25 +0800
  • c66bc35cfe Merge branch 'main' of https://github.com/hpcaitech/ColossalAI into jiaruifang/polish_sharded_model_v2 jiaruifang 2022-03-30 14:07:59 +0800
  • 92f4224867
    Automated submodule synchronization (#501) github-actions[bot] 2022-03-30 14:06:23 +0800
  • d74db1ad07 Merge branch 'main' of https://github.com/hpcaitech/ColossalAI into jiaruifang/polish_sharded_model_v2 jiaruifang 2022-03-30 14:03:19 +0800
  • 2e9c3e6cff polish jiaruifang 2022-03-30 14:01:48 +0800
  • 214da761d4
    [zero] add stateful tensor (#549) Jiarui Fang 2022-03-30 13:51:37 +0800
  • a569d1a7ce polish jiaruifang 2022-03-30 13:26:35 +0800
  • 71ba175ad0 polish jiaruifang 2022-03-30 12:07:59 +0800
  • 3077d027eb polish jiaruifang 2022-03-30 11:22:01 +0800
  • ca32ffc893 polish code jiaruifang 2022-03-30 11:16:46 +0800
  • d38b0f8a6f polish code jiaruifang 2022-03-30 11:06:43 +0800
  • e5224e7373 [zero] add stateful tensor jiaruifang 2022-03-30 10:58:16 +0800
  • 107b99ddb1
    [zero] dump memory stats for sharded model (#548) Jiarui Fang 2022-03-30 09:38:44 +0800
  • 1fd454122a polish jiaruifang 2022-03-30 09:37:51 +0800
  • 763dc325f1
    [TP] Add gather_out arg to Linear (#541) Ziyue Jiang 2022-03-30 09:35:46 +0800
  • 0da0c5ecdb Automated submodule synchronization github-actions 2022-03-30 00:01:06 +0000