Commit Graph

  • b1a9691201 online repack mrhaoxx 2026-02-24 11:01:53 +00:00
  • 950f052f96 ckpt mrhaoxx 2026-02-24 03:03:29 +00:00
  • 8aea4b3e2a ckpt mrhaoxx 2026-02-22 09:18:26 +00:00
  • cf09b3e1ca [fix]: skiplora mrhaoxx 2026-02-17 03:06:19 +00:00
  • 16a8b98f3e support qwen3.5 (#1846) Jianwei Dong 2026-02-16 15:48:14 +08:00
  • d23e7c4778 deploy: 411b69bec0 SkqLiao 2026-02-15 03:37:32 +00:00
  • 411b69bec0 Add GLM-5 Day0 Support update to README (#1851) Jiaqi Liao 2026-02-15 11:37:12 +08:00
  • 9e597a25a8 deploy: 7d9943365a SkqLiao 2026-02-13 14:49:35 +00:00
  • 7d9943365a Add MiniMax-M2.5 Day0 support update (#1850) Jiaqi Liao 2026-02-13 22:49:10 +08:00
  • a3d5d53605 Update MiniMax-M2.5.md (#1849) Jiaqi Liao 2026-02-13 22:35:36 +08:00
  • f0e4fc612b support minimax-m2.5 (#1848) Jianwei Dong 2026-02-13 19:15:44 +08:00
  • c631dbbd7a support minimax-m2.5 minimax-m2.5 djw 2026-02-13 11:05:26 +00:00
  • 1eafb7630a merge main qwen3.5 zhangxiaohao 2026-02-12 15:15:56 +08:00
  • 240237bd8f support qwen3.5 zhangxiaohao 2026-02-12 15:15:06 +08:00
  • 1c72b3f5bd fix glm5 docs (#1845) Oql 2026-02-12 02:33:37 +08:00
  • 7f7aeaeff6 support glm 5 (#1844) Oql 2026-02-12 02:03:32 +08:00
  • 061fb56382 Update Kimi-K2.5.md (#1838) Jiaqi Liao 2026-02-07 16:38:39 +08:00
  • d342fb1df6 [docs]: add maintainers list (#1837) Chen Hongtao 2026-02-06 18:28:06 +08:00
  • 7c3b74f493 Apply suggestion from @Copilot add-maintainers ErvinXie 2026-02-06 18:27:43 +08:00
  • 5bc9e64909 [docs]: add maintainers list chenht2022 2026-02-06 10:12:26 +00:00
  • a1a9eca311 [fix]: perf mrhaoxx 2026-02-05 16:57:09 +00:00
  • b0740436b5 [fix]: perf mrhaoxx 2026-02-05 14:52:59 +00:00
  • 8825cbf4b6 [fix]: perf mrhaoxx 2026-02-04 15:26:04 +00:00
  • 56cbd69ac4 kt-cli enhancement (#1834) Oql 2026-02-04 16:44:54 +08:00
  • 4eb1669c9a deploy: 4f64665758 ouqingliang 2026-02-04 08:27:23 +00:00
  • 4f64665758 [docs]: add Qwen3 Coder Next Tutorial (#1833) Oql 2026-02-04 16:27:10 +08:00
  • d36515d22d [docs]: support Qwen3 Coder Next Tutorial oql-doc ouqingliang 2026-02-04 08:25:37 +00:00
  • 3b2de00593 [fix]: wip mrhaoxx 2026-02-04 05:40:47 +00:00
  • fac81ed147 [fix]: fix tp mrhaoxx 2026-02-03 14:46:40 +00:00
  • 59c243d371 [fix]: fix cpu cores. cli-dev-ljq ouqingliang 2026-02-03 09:17:30 +00:00
  • 2bcc0c8cb5 Merge branch 'main' into cli-dev-ljq ouqingliang 2026-02-03 08:10:53 +00:00
  • c28cfcb26e [fix]: fix k2-moe.hpp load weight (#1830) Oql 2026-02-03 11:28:49 +08:00
  • 232617bd29 v0 djw 2026-02-02 09:06:45 +00:00
  • 391cb6f79d [fix]: fix memory footprint mrhaoxx 2026-02-02 08:34:01 +00:00
  • 06fb3b5dbf [fix]: prequant weight load mrhaoxx 2026-02-01 15:17:47 +00:00
  • 9efe1317b1 [fix]: fix forward cache (maybe) mrhaoxx 2026-01-31 21:19:46 +00:00
  • e1e64f7948 [fix]: fix lora grad compute mrhaoxx 2026-01-31 15:41:55 +00:00
  • 794c04fae4 Revert "[doc]: update kimi_k2.5 doc (#1823)" (#1825) Jiaqi Liao 2026-01-30 16:10:01 +08:00
  • ccbb5b1cf8 [docs]: add clawd bot docs Oql 2026-01-30 15:51:30 +08:00
  • 2e6506535b [doc]: update kimi_k2.5 doc (#1823) Jiaqi Liao 2026-01-30 15:43:18 +08:00
  • 630a2734a5 [feat]: improve CLI table display, model verification, and chat experience skqliao 2026-01-30 07:32:32 +00:00
  • b432e27d3a Merge branch 'main' into cli-dev-ljq skqliao 2026-01-30 07:32:03 +00:00
  • db82d99fa6 feat: add fallback expert prefix lookup in loader.py from kimi_k2.5 (#1822) Jiaqi Liao 2026-01-30 14:09:38 +08:00
  • 7f1b61e8e6 [docs]: 添加 Clawdbot 飞书接入教程链接 ouqingliang 2026-01-29 12:50:54 +00:00
  • 7d6961c805 [docs]: 强调推荐使用 Kimi K2.5 作为核心模型,突出企业级推理能力 ouqingliang 2026-01-29 12:45:33 +00:00
  • f5762391f2 [docs]: 添加 Clawdbot 集成指南 - KTransformers 企业级 AI 助手部署方案 ouqingliang 2026-01-29 12:39:35 +00:00
  • 5d0fdfd34c [fix]: correct TTFT measurement in kt chat skqliao 2026-01-28 12:00:26 +00:00
  • 8268d9aeae [feat]: improve kt chat with performance metrics and better CJK support skqliao 2026-01-28 11:56:24 +00:00
  • 2d343e055a [feat]: unify model selection UI and enable shared experts fusion by default skqliao 2026-01-28 11:36:28 +00:00
  • a052fc3809 [feat]: redesign kt run interactive configuration with i18n support skqliao 2026-01-28 11:22:46 +00:00
  • ee6ec0ac32 [feat]: merge from main ljq-cli-dev-new skqliao 2026-01-28 09:00:54 +00:00
  • edc48aba37 [fix]: fix wrapper import issue (#1819) Jiaqi Liao 2026-01-28 16:31:56 +08:00
  • 6cda1d5404 layerwise prefill opt djw 2026-01-27 09:55:23 +00:00
  • fa3b61da20 deploy: 8321d00cc5 JimmyPeilinLi 2026-01-27 09:45:11 +00:00
  • 8321d00cc5 Add files via upload (#1814) Peilin Li 2026-01-27 17:44:50 +08:00
  • f2747a4c88 Add files via upload JimmyPeilinLi-patch-2 Peilin Li 2026-01-27 17:43:57 +08:00
  • b941447a83 [fix]: fix import errors kimi_k2.5 SkqLiao 2026-01-27 16:11:52 +08:00
  • c7ce4c3e83 deploy: 2f6f7f1921 SkqLiao 2026-01-27 05:33:47 +00:00
  • 2f6f7f1921 Kimi k2.5 doc (#1812) Jiaqi Liao 2026-01-27 13:33:25 +08:00
  • 6a1e7c48cb [fix]: direct accumulation mrhaoxx 2026-01-27 04:27:59 +00:00
  • 1da075a3fa Revert "[doc]: add Kimi-K2.5 deploy&sft guide (#1810)" (#1811) Jiaqi Liao 2026-01-27 10:05:13 +08:00
  • a368140d76 [doc]: add Kimi-K2.5 deploy&sft guide (#1810) Jiaqi Liao 2026-01-27 10:02:59 +08:00
  • 014f060a83 [feat]: fix & update kimi_k2.5 SkqLiao 2026-01-27 09:54:43 +08:00
  • 7b62d826e4 [fix]: pinned memory causes numa issue mrhaoxx 2026-01-26 12:04:46 +00:00
  • 7b432f4b5a [fix]: avoid unnecessary memcpy mrhaoxx 2026-01-26 05:34:19 +00:00
  • 773ac20847 [fix]: fix missing bufferB init mrhaoxx 2026-01-25 18:58:49 +00:00
  • 192a9584f1 [chore]: save mrhaoxx 2026-01-25 18:28:40 +00:00
  • 63863b6322 [feat]: disable timer mrhaoxx 2026-01-25 17:43:40 +00:00
  • b53a3dbb2b [feat]: use vectorized transpose mrhaoxx 2026-01-25 17:31:35 +00:00
  • ae83d8237b [fix]: use buffer pool mrhaoxx 2026-01-25 17:09:31 +00:00
  • 32dfc5390c [feat]: merge some kernel mrhaoxx 2026-01-25 16:45:48 +00:00
  • 0669b910aa [feat]: optmize to use tr lora params mrhaoxx 2026-01-25 16:17:42 +00:00
  • 9abb104c9b [feat]: optimize kernels mrhaoxx 2026-01-25 13:00:29 +00:00
  • f80fe1682f [feat]: introduce json profiler mrhaoxx 2026-01-24 17:18:54 +00:00
  • 57580016ea [feat]: support async sft forward mrhaoxx 2026-01-24 14:45:08 +00:00
  • 03a710bc68 [chore]: remove some metrics mrhaoxx 2026-01-24 09:42:18 +00:00
  • 15d91e0880 [fix]: fix scheduling mrhaoxx 2026-01-23 19:04:30 +00:00
  • ae1252e874 [fix]: fix buffer A out of bounds read mrhaoxx 2026-01-23 18:10:08 +00:00
  • 451c91dce1 [fix]: nan mrhaoxx 2026-01-23 15:30:39 +00:00
  • a207f14411 deploy: 5bd5c8f750 ouqingliang 2026-01-23 10:25:29 +00:00
  • 6a97d94f78 deploy: bf4c8a690b ouqingliang 2026-01-23 10:19:09 +00:00
  • 72c04021a6 deploy: 5bd5c8f750 ouqingliang 2026-01-23 10:11:16 +00:00
  • c49ac26ec5 deploy: bf4c8a690b ouqingliang 2026-01-23 10:06:30 +00:00
  • 5bd5c8f750 [fix]: fix experts-sched-Tutorial.md (#1808) Oql 2026-01-23 18:06:24 +08:00
  • 978dedee75 [fix]: fix experts-sched-Tutorial.md oql-dev ouqingliang 2026-01-23 10:04:45 +00:00
  • bf4c8a690b Add Native Precision Tutorial, update worker strategy and README.md (#1807) Oql 2026-01-23 18:00:13 +08:00
  • 503d109fbc [fix]: fix buffer memory overuse mrhaoxx 2026-01-23 07:08:59 +00:00
  • d90f035735 [fix]: fix memory overflow mrhaoxx 2026-01-23 05:40:30 +00:00
  • 8652346e69 [fix]: doc (#1805) Jianwei Dong 2026-01-23 11:17:08 +08:00
  • 716794ac4c merge main experts_sched djw 2026-01-23 03:16:21 +00:00
  • b0f827d2a9 [chore](cuda): explicitly use ele_per_blk var for better readability (#1784) SCDESPERTATE 2026-01-23 11:11:08 +08:00
  • 898a69c887 fix doc djw 2026-01-23 03:08:06 +00:00
  • d50d19fcf9 [fix]: remove debug message mrhaoxx 2026-01-22 20:14:45 +00:00
  • 8f44a64a7a [fix]: fix memory pool mrhaoxx 2026-01-22 19:41:07 +00:00
  • 779bf14556 [doc]: add Experts sched tutorial (#1802) v0.5.1 Jianwei Dong 2026-01-22 15:40:07 +08:00
  • d4be481535 add experts_sched tutorial djw 2026-01-22 07:37:35 +00:00
  • 5c89cec5e3 [fix]: optimize job sched mrhaoxx 2026-01-22 07:04:52 +00:00
  • 23d4aabd14 deploy: a4de664e62 ovowei 2026-01-22 06:53:05 +00:00
  • a4de664e62 Add AutoDL Tutorial (#1801) Peilin Li 2026-01-22 14:52:47 +08:00
  • 8ff417f46c [feat]: vectorized lora compute mrhaoxx 2026-01-22 06:17:09 +00:00