Commit Graph

8086 Commits

Author SHA1 Message Date
Ke Bao
8fe3e37468 Support piecewise cuda graph for dsv3 fp4 (#15531) 2025-12-21 14:50:32 +08:00
YAMY
6014365564 [EAGLE] Fix slow Triton compilation in EAGLE KV cache copy by chunking large num_locs_upper (#15111) 2025-12-21 14:14:14 +08:00
Yinghai Lu
fc05acc2c7 [FusedMoE] Fix fused w13 tp sharded weight loading (#15432) 2025-12-21 13:59:36 +08:00
Mick
1ed946680a [diffusion] bench: improve bench_serving by adding more controlling args (#15554) 2025-12-21 13:37:45 +08:00
Mick
d7fbe73bf2 [diffusion] chore: minor improvements and typo-fixing (#15556) 2025-12-21 13:37:10 +08:00
Xiaoyu Zhang
42bff706df [diffusion] profiling: simplify --perf-dump-path JSON output (remove duplicate denoise steps) (#15537)
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-12-21 12:56:46 +08:00
Haoran Wang
26704c23c0 [model-gateway]: Tool parser for glm47 (#15520) 2025-12-20 18:07:51 -10:00
Alison Shao
43b7c17474 [CI] Fix /rerun-stage command by using requests for workflow dispatch (#15447) 2025-12-20 19:50:56 -08:00
Douglas Yang
96740d6983 feature: revamp nightly tests with combined runner (#15324) 2025-12-20 19:25:22 -08:00
Alison Shao
9a3bdf2c95 [CI] Migrate CUDA Graph tests to test/registered/cuda_graph/ (#15436) 2025-12-20 18:55:39 -08:00
Xiaoyu Zhang
4b351f6b95 Apply new moe align block size kernel (#14134) 2025-12-21 10:13:32 +08:00
Xiaoyu Zhang
7fa4906f4f [sgl-kernel] Streamline kernel size report (Top 20 only) and clean up (#15552) 2025-12-21 10:00:47 +08:00
Yuan Luo
050f108c29 Optimize Bailing-MoE with FlashInfer Fused All-Reduce (#15526)
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
2025-12-20 17:34:04 -08:00
Xinyuan Tong
47cdb65a45 fix: update argument extraction in R1 chat template (#15547)
Signed-off-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
2025-12-21 09:18:49 +08:00
shuwenn
1d90b194b2 [model-gateway] bugfix: backward compatibility for GET endpoints (#15413) 2025-12-20 10:37:53 -10:00
Praneth Paruchuri
537ef18d17 [model-gateway] Optimize WASM Runtime with Instance Pooling and Component Caching (#15515) 2025-12-20 10:27:02 -10:00
Simo Lin
69412ccb1f [model-gateway] add model gateway multi-arch docker build, test and document docker image (#15544) 2025-12-20 11:56:59 -08:00
Simo Lin
d3885d4ba6 Fix docker gateway image name and add latest tag (#15542) 2025-12-20 11:05:03 -08:00
Xinyuan Tong
0a346d3bd9 feat: Add limit-mm-data-per-request argument to server arguments (#15418)
Signed-off-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
2025-12-20 10:44:47 -08:00
Mick
bc18cb8631 [diffusion] refactor: change zmq socket type to router for scheduler (#15479) 2025-12-21 00:39:54 +08:00
Xiaoyu Zhang
bee8ac5b88 [diffusion] doc: add --perf-dump-path section to profiling doc (#15533) 2025-12-21 00:15:27 +08:00
Mick
41bd76e18b [diffusion] log: fix wrong use of suppress_other_loggers (#15534) 2025-12-20 23:43:55 +08:00
Mick
c6ca1b3afc [diffusion] chore: allow all attention backends if not specified (#15530) 2025-12-20 23:14:55 +08:00
Xiaoyu Zhang
8999ce754f [diffusion] perf: support zero-cost weight offload and overlap with compute for wan-series (#15511) 2025-12-20 22:52:40 +08:00
Yingchun Lai
dce2ed4467 fix MiMo-V2-Flash typo (#15536) 2025-12-20 22:49:06 +08:00
Yuan Luo
019517a356 [VLM] Support ViT Piecewise CUDA Graph for Qwen3-VL (#15320)
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
2025-12-20 21:00:07 +08:00
mlmz
1f1f05a85e vlm: refactor engine vlm params and support processor output as input (#14091)
Co-authored-by: Mick <mickjagger19@icloud.com>
Co-authored-by: zhaochenyang20 <zhaochenyang20@gmail.com>
Co-authored-by: Xinyuan Tong <115166877+JustinTong0323@users.noreply.github.com>
Co-authored-by: BenYao21 <cyao22@asu.edu>
Co-authored-by: minleminzui <minleminzui@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: 赵晨阳 <zhaochen20@outlook.com>
2025-12-20 18:31:24 +08:00
Yuan Luo
165f5c04cb Optimize MiMo-V2-Flash by flashinfer fused allreduce (#15464)
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
2025-12-20 17:45:18 +08:00
Артем Савкин
3e01f3a533 [NPU] [BUGFIX] [CRITICAL!] Fix NPU inference (torch_npu._npu_reshape_and_cache() crash) (#15484) 2025-12-20 14:53:46 +08:00
Hubert Lu
51e2eaa458 [AMD] Support fast_topk kernels in sgl-kernel (#15172) 2025-12-19 22:19:09 -08:00
yctseng0211
6468cb5823 [AMD] remove the redundant projection (#15178) 2025-12-19 22:10:46 -08:00
Qiaolin Yu
e220da1723 tiny fix sampling seed for completion api (#15498) 2025-12-19 20:32:13 -08:00
Yuxuan Zhang
b82c7a0ae7 [GLM-4.7] GLM-4.7 Tool Parser and Doc Update (#15333) 2025-12-19 20:30:44 -08:00
Mick
c0f9b51992 [diffusion] refactor: deprecate workloadtype (#15267)
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2025-12-20 11:24:31 +08:00
Simo Lin
5529ab5895 [model-gateway] Implement RAII load guard with response body attachment (#15507) 2025-12-19 19:14:52 -08:00
Simo Lin
74a3349bea [ci] remove rust benchmark in unit test ci (#15510) 2025-12-19 18:59:41 -08:00
Liangsheng Yin
b9ebf0ed63 Clean hidden_states_before_norm (#15485) 2025-12-20 09:38:46 +08:00
kun-llfl
3c116d5e5a [router] bugfix: cache_aware in grpc inbalance forward (#15473)
Signed-off-by: Kun(llfl) <i@imux.top>
2025-12-19 16:25:46 -08:00
Simo Lin
71a602883a [model-gateway] simplify workflow engine backoff and reduce duplicate reads (#15505) 2025-12-19 16:23:02 -08:00
Simo Lin
61405b3d00 [model-gateway] Run workflow event subscribers concurrently (#15504) 2025-12-19 15:49:51 -08:00
Simo Lin
0adfc42b99 [model-gateway] Optimize workflow engine with pre-computed dependency graph (#15503) 2025-12-19 15:44:57 -08:00
Simo Lin
ba72e759ca [model-gateway] Improve logging across core modules (#15497) 2025-12-19 13:35:28 -08:00
YL D
6afc5d497b [DSv32] Move deep_gemm.get_paged_mqa_logits_metadata to init time as metadata (#15040) 2025-12-19 13:23:33 -08:00
michael-amd
2ee6c810b8 [AMD] Add TP=8 models to nightly test and make TP=2 test stable (#15296) 2025-12-19 13:15:19 -08:00
Simo Lin
bd16244d93 [model-gateway] Improve logging in policies module (#15496) 2025-12-19 13:02:57 -08:00
Simo Lin
b5eb0214ca [model-gateway] Improve logging in data_connector module (#15495) 2025-12-19 12:57:18 -08:00
Simo Lin
d72e908ba5 [model-gateway] refactor: extract common graceful shutdown code before TLS branch (#15494) 2025-12-19 12:56:33 -08:00
Simo Lin
50cad014dc [model-gateway] fix graceful shutdown for TLS/Non-TLS server (#15491) 2025-12-19 12:35:50 -08:00
Khush Gupta
ef908aeb40 fixed trtllm nvfp4 backend for moe (#15022) 2025-12-19 10:49:21 -08:00
Yuwei An
9d0347b33a EP Support for Piecewise Cuda Graph (#14164)
Signed-off-by: Oasis-Git <ayw.sirius19@gmail.com>
2025-12-20 01:59:27 +08:00