Commit Graph

  • 3420921101 [shardformer] DeepseekMoE support (#5871) Haze188 2024-07-05 16:13:58 +08:00
  • e17f835df7 [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2024-07-04 12:47:16 +00:00
  • 6991819a97 Merge branch 'hpcaitech:main' into feature/fp8_comm Hanks 2024-07-04 20:34:41 +08:00
  • 7997683aac [pre-commit.ci] pre-commit autoupdate (#5878) pre-commit-ci[bot] 2024-07-04 13:46:41 +08:00
  • 7afbc81d62 [quant] fix bitsandbytes version check (#5882) Hongxin Liu 2024-07-04 11:33:23 +08:00
  • 6cd4c32be4 [shardformer] fix the moe (#5883) Wang Binluo 2024-07-03 20:02:19 +08:00
  • eb24fcd914 [Hotfix] Fix OPT gradient checkpointing forward Edenzzzz 2024-07-03 14:57:57 +08:00
  • ea94c07b95 [hotfix] fix the bug that large tensor exceed the maximum capacity of TensorBucket (#5879) Haze188 2024-07-02 12:42:02 +08:00
  • 7c2f79fa98 [pre-commit.ci] pre-commit autoupdate (#5572) pre-commit-ci[bot] 2024-07-01 17:16:41 +08:00
  • 936d0b0f7b [doc] Update llama + sp compatibility; fix dist optim table Edenzzzz 2024-07-01 17:07:22 +08:00
  • 8ab46b4000 [Shardformer] change qwen2 modeling into gradient checkpointing style (#5874) Jianghai 2024-07-01 16:45:09 +08:00
  • f5a52e1600 fp8 operators for compressed communication HangXu 2024-07-01 13:44:21 +08:00
  • ff535204fe update transformers version YeAnbang 2024-06-28 06:24:30 +00:00
  • 416580b314 [MoE/ZeRO] Moe refactor with zero refactor (#5821) Haze188 2024-06-28 14:00:08 +08:00
  • a8af6ccb73 fix torch colossalai version YeAnbang 2024-06-28 03:58:29 +00:00
  • 773d9f964a [shardformer]delete xformers (#5859) flybird11111 2024-06-28 11:20:04 +08:00
  • e7527762a1 Merge branch 'main' of https://github.com/hpcaitech/ColossalAI into rlhf_SimPO YeAnbang 2024-06-28 02:50:14 +00:00
  • eaea88cf9e [release] update version (#5864) v0.4.0 Hongxin Liu 2024-06-28 10:49:55 +08:00
  • 3c7cda0c9a [Inference]Lazy Init Support (#5785) Runyu Lu 2024-06-27 18:02:15 +08:00
  • d9d5e7ea1f [shardformer] Support the T5ForTokenClassification model (#5816) Guangyao Zhang 2024-06-27 16:40:38 +08:00
  • 5dfbcd7746 [zero] use bucket during allgather (#5860) Hongxin Liu 2024-06-27 16:34:44 +08:00
  • b117274074 fix colossalai, transformers version YeAnbang 2024-06-27 08:30:17 +00:00
  • afa53066ca fix colossalai, transformers version YeAnbang 2024-06-27 08:28:36 +00:00
  • 384c64057d fix colossalai, transformers version YeAnbang 2024-06-27 08:26:44 +00:00
  • 8aad064fe7 fix style YeAnbang 2024-06-27 07:29:33 +00:00
  • c8d1b4a968 add orpo YeAnbang 2024-06-27 07:20:28 +00:00
  • 868afdb311 Dev/zero offload (#5858) dev/zero-offload Wang Binluo 2024-06-26 16:07:06 +08:00
  • 8e718a1421 [gemini] fixes for benchmarking (#5847) botbw 2024-06-26 15:52:09 +08:00
  • 2a25a2aff7 [Feature] optimize PP overlap (#5735) Edenzzzz 2024-06-26 14:48:02 +08:00
  • de3f67d128 fix llama (#5856) Wang Binluo 2024-06-26 10:15:13 +08:00
  • 4ccaaaab63 [doc] add GPU cloud playground (#5851) binmakeswell 2024-06-25 11:03:16 +08:00
  • f3de5a025c remove debug code YeAnbang 2024-06-24 05:16:29 +00:00
  • 0b2d6275c4 fix dataloader YeAnbang 2024-06-24 05:10:44 +00:00
  • 4b59d874df Merge branch 'main' of https://github.com/hpcaitech/ColossalAI into main YeAnbang 2024-06-24 02:16:03 +00:00
  • 82aecd6374 add SimPO colossalchat_upgrade YeAnbang 2024-06-24 02:12:20 +00:00
  • 7266f82d03 [doc] fix open sora model weight link (#5848) binmakeswell 2024-06-21 22:48:34 +08:00
  • 8f445729a4 [doc] opensora v1.2 news (#5846) binmakeswell 2024-06-21 14:20:45 +08:00
  • 8a5c86439a [gemini] fix missing return (#5845) botbw 2024-06-21 11:38:40 +08:00
  • 4c06215dce Merge pull request #5844 from wangbluo/offload Wang Binluo 2024-06-20 17:07:57 +08:00
  • e893f88a4f Merge branch 'dev/zero-offload' into offload Wang Binluo 2024-06-20 17:07:24 +08:00
  • d4ff644ef3 update qwen model wangbluo 2024-06-20 09:04:57 +00:00
  • dba59354d7 remove vocab_size args wangbluo 2024-06-20 08:06:39 +00:00
  • 35ef72bfd1 Merge pull request #5842 from wangbluo/dev/zero-offload Wang Binluo 2024-06-20 15:37:03 +08:00
  • 351a1c269b [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2024-06-20 06:50:39 +00:00
  • b12e9a3275 update llama model wangbluo 2024-06-20 06:46:25 +00:00
  • bd3e34fef6 [release] update version (#5833) v0.3.9 Hongxin Liu 2024-06-20 13:33:24 +08:00
  • 52ea64824e remove 4d attention mask wangbluo 2024-06-19 09:28:08 +00:00
  • 7b249c76e5 [Fix] Fix spec-dec Glide LlamaModel for compatibility with transformers (#5837) Yuanheng Zhao 2024-06-19 15:37:53 +08:00
  • fd1dc417d8 [shardformer] Change atol in test command-r weight-check to pass pytest (#5835) Guangyao Zhang 2024-06-19 13:59:22 +08:00
  • 2014cce870 [devops] Remove building on PR when edited to avoid skip issue (#5836) Guangyao Zhang 2024-06-19 13:58:05 +08:00
  • 0adca5b688 [launch] Support IPv4 host initialization in launch (#5822) Kai Lv 2024-06-18 19:18:29 +08:00
  • 639394b0d4 Merge pull request #5818 from GuangyaoZhang/command-r Guangyao Zhang 2024-06-18 19:01:21 +08:00
  • 7f9ec599be [misc] Add dist optim to doc sidebar (#5806) Edenzzzz 2024-06-18 13:52:47 +08:00
  • 4adbc36913 Merge branch 'command-r' of github.com:GuangyaoZhang/ColossalAI into command-r GuangyaoZhang 2024-06-18 03:33:02 +00:00
  • d84d68601a change 'xxx if xxx else None' to 'xxx or None' GuangyaoZhang 2024-06-18 03:32:42 +00:00
  • 996c65077e [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2024-06-18 03:32:29 +00:00
  • a83a2336e8 rebase master llama change GuangyaoZhang 2024-06-18 02:56:47 +00:00
  • 20c0b06ff5 Merge branch 'command-r' of github.com:GuangyaoZhang/ColossalAI into command-r GuangyaoZhang 2024-06-18 02:37:14 +00:00
  • 363cde6957 merge model and attention forward GuangyaoZhang 2024-06-17 08:50:05 +00:00
  • 7a2b08646f Remove CohereLayerNorm and use existing layernorm GuangyaoZhang 2024-06-14 09:14:01 +00:00
  • fe2e74c03a fix precommit GuangyaoZhang 2024-06-14 08:09:24 +00:00
  • 98da648a4a Fix Code Factor check GuangyaoZhang 2024-06-14 08:04:29 +00:00
  • f656d61778 change command GuangyaoZhang 2024-06-14 03:04:56 +00:00
  • 0b81163bc0 Copy llama to command GuangyaoZhang 2024-06-13 06:47:58 +00:00
  • 8795bb2e80 Support 4d parallel + flash attention (#5789) Edenzzzz 2024-06-17 17:40:47 +08:00
  • 3c7302ad0e merge model and attention forward GuangyaoZhang 2024-06-17 08:50:05 +00:00
  • 8c3f524660 Remove CohereLayerNorm and use existing layernorm GuangyaoZhang 2024-06-14 09:14:01 +00:00
  • df612434c9 [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2024-06-14 05:02:37 +00:00
  • 4c69e2dc91 support qwen model Wang Binluo 2024-04-09 11:50:35 +08:00
  • 32e642bf40 revert: enable return_outputs when necessary Wenhao Chen 2024-03-01 16:50:58 +08:00
  • 856b39f69d to: add qwen2 auto policy Wenhao Chen 2024-03-01 16:55:04 +08:00
  • 6fa181ebef feat: add qwen2 to model_zoo Wenhao Chen 2024-03-01 17:23:46 +08:00
  • 14305c9449 test: add qwen2 shard test Wenhao Chen 2024-03-01 17:26:20 +08:00
  • 5512bdf1fc fix: modify model config and add Qwen2RMSNorm Wenhao Chen 2024-03-04 11:22:34 +08:00
  • 5c2a47a667 feat: support qwen2 model Wenhao Chen 2024-03-04 13:33:11 +08:00
  • c9025ebd7c Merge branch 'command-r' of github.com:GuangyaoZhang/ColossalAI into command-r GuangyaoZhang 2024-06-14 08:10:31 +00:00
  • 9a290ab013 fix precommit GuangyaoZhang 2024-06-14 08:09:24 +00:00
  • 2a7fa2e7d0 [pre-commit.ci] auto fixes from pre-commit.com hooks pre-commit-ci[bot] 2024-06-14 08:05:06 +00:00
  • 1016bb3257 Fix Code Factor check GuangyaoZhang 2024-06-14 08:04:29 +00:00
  • 94fbde6055 change command GuangyaoZhang 2024-06-14 03:04:56 +00:00
  • 431b7bcf8f Copy llama to command GuangyaoZhang 2024-06-13 06:47:58 +00:00
  • 2ddf624a86 [shardformer] upgrade transformers to 4.39.3 (#5815) flybird11111 2024-06-14 10:59:33 +08:00
  • 3bcbba9262 [gemini] quick fix on possible async operation (#5803) botbw 2024-06-13 10:35:17 +08:00
  • d9dddf574f [Gemini] Use async stream to prefetch and h2d data moving (#5781) Haze188 2024-06-12 15:48:52 +08:00
  • 8554585a5f [Inference] Fix flash-attn import and add model test (#5794) Li Xingjian 2024-06-12 14:13:50 +08:00
  • aac941ef78 [test] fix qwen2 pytest distLarge (#5797) Guangyao Zhang 2024-06-12 12:13:51 +08:00
  • aa125bcc91 [shardformer] fix modeling of bloom and falcon (#5796) Hongxin Liu 2024-06-11 17:43:50 +08:00
  • 587bbf4c6d [test] fix chatglm test kit (#5793) Hongxin Liu 2024-06-11 16:54:31 +08:00
  • 74f4a29734 Merge pull request #5759 from hpcaitech/colossalchat_upgrade YeAnbang 2024-06-11 12:49:53 +08:00
  • c0948aff97 [Inference]refactor baichuan (#5791) Runyu Lu 2024-06-11 10:52:01 +08:00
  • 84eab13078 update sft trainning script YeAnbang 2024-06-11 02:44:20 +00:00
  • 77a219a082 Merge pull request #5771 from char-1ee/refactor/modeling Li Xingjian 2024-06-10 11:52:22 +08:00
  • b303976a27 Fix test import char-1ee 2024-06-10 02:03:30 +00:00
  • 2abdede1d7 fix readme YeAnbang 2024-06-10 01:08:42 +00:00
  • f5981e808e Remove flash attention backend char-1ee 2024-06-07 10:02:19 +00:00
  • 77db21610a replace the customized dataloader setup with the build-in one YeAnbang 2024-06-07 09:44:25 +00:00
  • 0d7ff10ea5 replace the customized dataloader setup with the build-in one YeAnbang 2024-06-07 09:43:42 +00:00
  • ceba662d22 Clean up char-1ee 2024-06-07 09:09:29 +00:00
  • 5f398fc000 Pass inference model shard configs for module init char-1ee 2024-06-07 08:28:19 +00:00
  • eec77e5702 Fix tests and naming char-1ee 2024-06-03 05:41:32 +00:00