70 Commits
Author SHA1 Message Date
Jiajun Li 1647bd3b25 Bump sglang to v0.5.22 (#4125) 2026-10-16 17:43:49 -07:00
Zhichen Zeng 8f409c7c49 fix(docker): patch TileLang's fp16 256-bit store on Blackwell (#3647) 2026-09-30 12:11:13 -07:00
Zhichen Zeng cc76e23915 GLM 5.3 Flash RL Support (#2786) 2026-09-24 20:05:11 -07:00
HEJIAN SANG b3c8f9c8f2 docker: Upgrade TileLang to 0.1.14 (#3309) 2026-09-22 17:10:58 -07:00
Yueming YuanandYueming Yuan 66b39f07a1 Bump TransformerEngine to 2.17.0 (#1781)
Co-authored-by: Yueming Yuan <yueming@Yuemings-MacBook-Air.local>
2026-07-24 00:01:18 -07:00
Jiajun Li ec6201de84 [docker] Default x86 image to CUDA 13 (sglang v0.5.12, cu130) (#1301) 2026-06-08 16:09:50 -07:00
Kaixi HouandJiajun Li a0b75e7f19 [NVIDIA] B300 (x86_64, sm_103a) Container Support (#977)
Co-authored-by: Jiajun Li <guapisolo@gmail.com>
2026-05-15 18:58:45 -07:00
Jiajun LiandClaude Opus 4.6 24627f0697 [cleanup] Remove expired docker patches to avoid confusion (#787)
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-24 23:53:49 -07:00
Jiajun Li 4844eee839 [fix] trivial fix lint (#786) 2026-03-22 14:55:37 -07:00
1bf014b15e [feat] GB300 (sm103a / cu13 arm64) support (#638)
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-authored-by: Ethan (Yusheng) Su <yushengsu.thu@gmail.com>
2026-03-21 18:52:33 -07:00
Yueming Yuan 4ebc75e4da [docker, CI] use radixark/Megatron-LM' and allow specify sglang/megatron PR in CI (#620) 2026-02-24 20:29:20 -08:00
Jiajun Li fa8458847a [fix] bypass r3 for mtp layer. (#619) 2026-02-19 00:17:18 -08:00
Yueming Yuanandfzyzcjy 6d15e48a73 Refactor and fix rollout routing replay (R3) (#575)
Co-authored-by: fzyzcjy <ch271828n@outlook.com>
2026-02-17 16:18:35 -08:00
Yueming Yuan 8133a6f0e2 [docker] upload dockerfile using sglang-miles branch and wheel (#582) 2026-02-13 14:34:04 -08:00
Jiajun Li 32d4c8552d feat: Support OAI TITO v1 (#502) 2026-02-05 22:35:48 +00:00
赵晨阳 20ab4f24e8 [squashed] Support VLM Multi-turn Training with Megatron, Support INT4 training, bug fix, etc. (#426) 2026-01-12 22:41:17 -08:00
赵晨阳 5d7a21c3e1 feat: add int4 reinforcement learning training support (Part1) (#422) 2026-01-12 17:06:21 -08:00
miles-code-angel 236b640bcc update code (#411) 2026-01-07 21:48:25 -08:00
miles-code-angelandRatish1 c77bdcf2a7 update code (#401)
Co-authored-by: Ratish1 <formula733@gmail.com>"
2026-01-05 13:32:45 -08:00
miles-code-angel 9dfd339247 update code (#385) 2026-01-03 20:09:56 -08:00
miles-code-angel 8a988dc3d2 [auto-sync] update code (#383) 2026-01-02 18:42:42 -08:00
miles.pr.bot 61a319ae0b update code 2025-12-13 13:37:21 +08:00
miles.pr.bot 67ba097a72 update code 2025-12-10 22:27:02 +08:00
miles.pr.bot f64d3ebbf6 update code 2025-11-29 21:25:18 +08:00
miles.pr.bot 9b14afe193 update code 2025-11-17 18:04:10 +08:00
fzyzcjy 18ff16ec0f Apply megatron commit d8c6aa4c to fix Blackwell missing saving checkpoint (#181) 2025-11-17 17:21:32 +08:00
miles.pr.bot d636ba5a05 update code 2025-11-17 11:40:26 +08:00
miles.pr.bot 3a1692e22a update code 2025-11-15 10:25:45 +08:00
fzyzcjy 65cbbc56bf Super tiny remove unused docker-related files (#151) 2025-11-14 17:30:07 +08:00
miles.pr.bot 350c5ee442 update code 2025-11-14 09:30:03 +08:00
miles.pr.bot a9ad745a9a update code 2025-11-01 17:06:56 +08:00
miles-bot 76456682b0 update code 2025-10-24 10:41:05 +08:00
miles-bot e0aa55f111 update code 2025-10-24 08:32:55 +08:00
miles-bot 86c8352a44 update code 2025-10-15 21:18:20 +08:00
miles-bot 7a33cb68d2 update code 2025-10-13 07:33:18 +08:00
Zilin Zhu c143ca2b23 [docker] patch offload bugfix on Qwen3Next in sglang 2025-10-04 14:44:06 +00:00
Zilin Zhu ab13a4aed8 [model] Support Qwen3-Next-80B-A3B (#417)
* [model] Support Qwen3-Next-80B-A3B

* bugfix

* use fla's ShortConvolution and support data packing
2025-10-04 22:12:07 +08:00
Zilin Zhu f291a8f5a5 add ad-hoc wait for large moe 2025-09-30 11:07:56 +00:00
Zilin Zhu eb5604e20f [docker] fix routing replay with pp and other bugfixes (#395) 2025-09-28 14:55:24 +08:00
Zilin Zhu 5b77a9b2f5 [feat] add --use-routing-replay (#387) 2025-09-26 20:38:56 +08:00
Zilin Zhu 69734956e1 [reproducibility][docker] enable training reproducibility (#370)
* [reproducibility][docker] enable training reproducibility

* bugfix

* bugfix

* add example

* update docs
2025-09-23 20:20:02 +08:00
Cheng-Yuan Lee 4d83c7c326 added gb200 patches (#360) 2025-09-22 15:22:21 +08:00
Zilin Zhu 4badb749b2 [feat] init support for PPO (#342) 2025-09-17 08:13:22 +08:00
maocheng23 b7938c83f6 add dockerfile and patch for b200 (#340)
* add dockerfile and patch for b200

* pre-commit for other files
2025-09-16 13:19:52 +08:00
Zilin Zhu 1099805fe3 [docker] update to sglang 0.5.2rc2 (#313) 2025-09-09 15:19:41 +08:00
Zilin Zhu 81c1f8661a [docker] bugfix for sglang dp attention 2025-09-02 22:59:04 -07:00
Jiajun Li 11f6fdef65 Update speculative decoding doc and sglang patch (#250)
* update doc for eagle and upd scripts

* upd sglang patch
2025-09-03 11:53:16 +08:00
Zilin Zhu e35f515758 [docker] set 0.5.0rc0 as new stable release 2025-08-29 00:22:25 -07:00
Zilin Zhu 771e0f2504 [docker] fix megatron patch 2025-08-26 00:05:53 -07:00
Yuzhen Zhou 149c348d4e fix: Add FSDP module backward compatibility to megatron.patch (#226)
* fix: Add FSDP module backward compatibility to megatron.patch

* fix pre-commit
2025-08-26 14:43:47 +08:00