Commit Graph
16997 Commits
Author SHA1 Message Date
MickandClaude Opus 5 0be2a209ac [diffusion] refactor: hand out pinned host memory per layer (#35867)
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
2026-08-22 09:37:31 +08:00
Mick b26695a26e [diffusion] feat: reject unsupported quantized component checkpoints (#35873) 2026-08-22 09:33:43 +08:00
MickandClaude Opus 5 22dafbcbd9 [diffusion] feat: keep a cpu-started vae weights on the checkpoint mapping (#35862)
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
2026-08-22 09:31:30 +08:00
Khoa PhamandClaude Opus 5 0ee3749e07 [Fix] Read the granite sinks dtype from the exec bag, not the legacy global shim (#35921)
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-21 18:13:18 -07:00
DarkSharpnessandClaude Opus 5 7fd5454335 [DSA] Route the ragged prefill top-k to the v2 kernel (#35175)
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-21 16:59:41 -07:00
Baizhou Zhang 60ff1e33a5 [DeepSeek V4] Default FP4 checkpoints to FlashInfer MXFP4 MoE (#35919) 2026-08-21 16:48:58 -07:00
Rain Jiang 7d7ab4b5c6 Rainj me/rust server refactor2 (#35239) 2026-08-21 16:37:02 -07:00
zijiexiaandClaude Opus 5 fe8f9d7457 [Docs] Add --prerelease=allow to cookbook uv install commands (#35920)
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
2026-08-21 15:51:48 -07:00
Alison Shao 2440820528 docs: add website link to README header (#35210) 2026-08-21 15:50:52 -07:00
YAMY 834400705f perf: overlap Qwen shared expert with DeepEP routed experts (#34938) 2026-08-21 15:39:44 -07:00
Yanbin Jiang 7d893255c3 [Spec][LoRA] Support multi-adapter LoRA with EAGLE/NEXTN/DFLASH/DSPARK speculative decoding (#34337) 2026-08-21 14:21:53 -07:00
Hank HanandTRAE CLI 590b11a5ef [Runtime] Don't override CUDA_MODULE_LOADING (#35711)
Co-authored-by: TRAE CLI <traecli@bytedance.com>
2026-08-21 14:09:27 -07:00
sglang-botandsglang-bot 4d42deff0a chore: bump docs install version to 0.5.18 (#35911)
Co-authored-by: sglang-bot <sglang-bot@users.noreply.github.com>
2026-08-21 13:24:03 -07:00
Xinyuan Tong 05c584c44f docs: add DSPARK speculative decoding option to Ling-3.0-flash cookbook (#35861) 2026-08-22 02:50:55 +08:00
Shangming Cai 729a050ea3 refactor(disagg): extract _all_reduce_polls helper (#35886) 2026-08-22 02:30:33 +08:00
Connor Carpenter c3735625de fix(grpc): derive choice count before normalization (#35778)
Signed-off-by: Connor Carpenter <connorc@nvidia.com>
2026-08-21 11:29:52 -07:00
Mohammad Miadh AngkadandMohammad Angkad 61c2da42bb [Fix] Pass Anthropic thinking history as reasoning_content for custom chat encoders (#35480)
Co-authored-by: Mohammad Angkad <mohammad.angkad@radixark.ai>
2026-08-22 02:22:11 +08:00
Davis Wertheimer 70983bd7db Add SGLang Granite SWA support via existing Granite models (#35794)
Signed-off-by: Davis Wertheimer <davis.wertheimer@ibm.com>
2026-08-21 11:13:15 -07:00
Shuwen Wangandispobock 0bdd28d487 [mem_cache] docs: add a layer map and placement rules (#35643)
Co-authored-by: ispobock <ispobaoke@gmail.com>
2026-08-22 00:35:07 +08:00
Shuwen Wangandispobock a7ec6b97f7 Restructure mem_cache auto-labels by layer (#25122)
Co-authored-by: ispobock <ispobaoke@gmail.com>
2026-08-21 23:29:59 +08:00
Mick 8658d00764 [diffusion] feat: support loading peft lora (#35868) 2026-08-21 22:58:59 +08:00
li_maxandMick 0447ade326 [diffusion] fix: fall back to a component's default attention backend (#35796)
Co-authored-by: Mick <mickjagger19@icloud.com>
2026-08-21 22:56:44 +08:00
王鹤男 932f632158 [diffusion] fix: do not warn that the recommended short edge is unverified (#35745) 2026-08-21 22:56:03 +08:00
Shangming Cai a41da991c8 refactor(disagg): collapse duplicated branches in get_kv_class (#35847) 2026-08-21 22:11:31 +08:00
王鹤男 5ecd6d794d [diffusion] fix: fix quantized qkv scales and missing-param policy for minimax-h3 (#35740) 2026-08-21 22:10:50 +08:00
Shangming Cai 4f343abc13 refactor(disagg): remove unreferenced dead code (#35838) 2026-08-21 22:09:25 +08:00
Mick 5a46d657b7 [diffusion] refactor: resolve lora weight sources deterministically (#35774) 2026-08-21 21:05:40 +08:00
Mick 5206f11543 [diffusion] fix: stop the mapped-weight store from holding the parameter itself (#35813) 2026-08-21 21:02:21 +08:00
Shangming CaiandClaude Opus 5 dad6fd0f04 refactor(disagg): remove dead get_embedding_port (#35844)
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-21 18:12:23 +08:00
Xiaoyu Zhang 39d4d65a51 [diffusion] Accelerate SANA-Video linear attention in quality=high (#35728) 2026-08-21 18:05:43 +08:00
Xiaoyu Zhang a5c52a9358 [diffusion] Enable LongCat breakable CUDA graphs (#35724) 2026-08-21 17:59:16 +08:00
Shangming Cai e7a37c8550 refactor(disagg): remove dead build_and_send_encode_request (#35843) 2026-08-21 17:58:35 +08:00
jasonjk-park 0db2c53dec Fix overlap prebuilt row reuse race (#35748) 2026-08-21 02:00:13 -07:00
Liangsheng Yin 896acc8860 [Fix] Clear full-to-SWA mapping with index_fill_ to avoid a blocking H2D copy (#35773) 2026-08-21 01:23:26 -07:00
4c98759c73 [AMD] fix(rocm): support flydsl 0.3.0 in the FlyDSL fused norm kernel (#34536)
Co-authored-by: Bingxu Chen <195740905+bingxche@users.noreply.github.com>
Co-authored-by: thomawan <thomawan@amd.com>
2026-08-21 01:13:21 -07:00
Shuwen Wang 8ff9c2b227 [mem_cache][9/N] refactor: move DSAIndexerPoolHost to pool_host.dsa (#35306) 2026-08-21 15:59:15 +08:00
8a123cbd0e [Refactor] New EPD (#30398)
Co-authored-by: Yuang Chen <1131578721@qq.com>
Co-authored-by: Yuang Chen <cya539102@antgroup.com>
Co-authored-by: ZhengWG <zwg0606@gmail.com>
2026-08-21 15:22:48 +08:00
Bingxu Chen 6a12583679 [AMD] Update ROCm AITER pin to c16d44b (#35810) 2026-08-21 00:02:17 -07:00
Zaili Wang f3fe81583e add py env activate in xpu kernel release workflow (#35726) 2026-08-20 23:27:00 -07:00
44c90c6282 [AMD] DSv4: fuse the qk-norm-rope pair on the MTP target-verify path (#34973)
Co-authored-by: HAI <hixiao@gmail.com>
Co-authored-by: Thomas Wang <thomawan@amd.com>
2026-08-20 23:08:22 -07:00
a688682f4b [AMD][CI] Fix ROCm 7.0's dead apt index fail the MORI dependency install (#35764)
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: quitenode <quitenode@users.noreply.github.com>
2026-08-20 22:13:09 -07:00
Ke Bao 73a2c117c6 Support mxfp8 KV cache in PD transfer (#35718) 2026-08-21 13:05:06 +08:00
Michaelandquitenode f64080fbaf [AMD] CI: cut two setup cycles from the AMD multimodal-gen lanes (#34483)
Co-authored-by: quitenode <quitenode@users.noreply.github.com>
2026-08-20 22:01:51 -07:00
Jimmy Shong 3efa057449 [docs] Retune the Qwen3.8-27B RTX 5090 DFLASH2 cells against 1cf2b8c (#35786) 2026-08-20 21:54:51 -07:00
Bingxu Chen 78c964d9d7 [AMD] Retry transient network failures in ROCm Dockerfile curl fetches (#35654) 2026-08-20 21:54:24 -07:00
Thomas Wang 34180a0d35 [AMD] Improve K3 dspark draft attn kernel perf (#35499) 2026-08-20 21:48:59 -07:00
bda9952377 [AMD] DeepSeek-V4 MI355X: eliminate bpreshuffle fp8-scale copies at producer sites (MoE down, MLA o_proj bmm) (#33166)
Co-authored-by: kk <43161300+kkHuang-amd@users.noreply.github.com>
Co-authored-by: Thomas Wang <thomawan@amd.com>
2026-08-20 21:46:23 -07:00
Mick 6127d1daee [diffusion] feat: allow offloaded weights stay on the checkpoint mapping (#35701) 2026-08-21 10:54:28 +08:00
Zhangheng 44806dc507 Using unified radix tree by default for all case (#35081) 2026-08-21 10:45:46 +08:00
MickandClaude Opus 5 e0cf75d9bd [doc] standardize diffusion cookbook model pages (#34247)
Co-authored-by: Claude Opus 5 <noreply@anthropic.com>
2026-08-21 10:25:40 +08:00