Commit Graph
15342 Commits
Author SHA1 Message Date
Woosuk Kwon d7ccecd2b7 Enable PW CUDA graphs
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-04-06 00:35:01 +00:00
Woosuk Kwon 1fed50d74f silu_mul_quant_fp4 fusion
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-04-03 01:05:18 +00:00
Woosuk Kwon f9bf662e5b fuse parallel linears
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-04-01 05:47:41 +00:00
Woosuk Kwon 14e2241f77 remove torch compile decorator
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-31 21:06:51 +00:00
Woosuk Kwon cdd23258cf dummy loader hack
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-31 19:25:36 +00:00
Woosuk Kwon cec6774e9b registry
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-31 19:23:24 +00:00
Woosuk Kwon 355be167e6 fuse & parallelize linears
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-31 06:30:37 +00:00
Woosuk Kwon d67e21b26e fuse moe scale_and_add
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-31 04:47:26 +00:00
Woosuk Kwon 6a2c13a6f0 fuse concat + fp8
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-31 03:59:16 +00:00
Woosuk Kwon b443e6702e fuse mla cache
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-31 03:38:42 +00:00
Woosuk Kwon 34d73a3375 Fuse indexer k cache update
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-30 22:28:33 +00:00
Woosuk Kwon 9d7beab915 indexer_k_quant_triton
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-30 19:47:36 +00:00
Woosuk Kwon f2ecfa9cd7 topk init
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-30 17:12:27 +00:00
Woosuk Kwon e4cdaf199d minor
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-30 17:02:23 +00:00
Woosuk Kwon 2b72935629 inline indexer
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-30 16:47:09 +00:00
Woosuk Kwon 1903df8328 all reduce + residual + rms norm
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-30 05:28:13 +00:00
Woosuk Kwon d872b0a082 fix
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-30 04:01:19 +00:00
Woosuk Kwon 84deceffb7 remove attn
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-30 03:07:54 +00:00
Woosuk Kwon e269b614c0 fuse quant & indexer weight update
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-30 03:04:25 +00:00
Woosuk Kwon 24090c52f3 adjust fused_q grid
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-30 02:31:20 +00:00
Woosuk Kwon 063fd29c98 fused op 1
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-30 01:49:10 +00:00
Woosuk Kwon 156e12ba35 indexer rope
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-29 23:11:34 +00:00
Woosuk Kwon 3e5c06dd7d rope
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-29 22:56:32 +00:00
Woosuk Kwon cc08dad785 layernorm
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-29 19:26:57 +00:00
Woosuk Kwon 976293e374 wip
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-29 17:26:23 +00:00
Woosuk Kwon 6efd919548 wip
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-29 17:19:31 +00:00
Woosuk Kwon 2145abaade Merge branch 'main' into woosuk/ds-exp 2026-03-27 22:23:32 +00:00
Giancarlo DelfinandGitHub 384e4d5f48 [Model Runner V2] Rebuild attention metadata before eagle decode full… (#38311)
Signed-off-by: Giancarlo Delfin <gdelfin@inferact.ai>
2026-03-27 13:46:42 -07:00
Nicolò LucchesiandGitHub 44a6528028 [CI] Skip failing test (#38369)
Signed-off-by: NickLucche <nlucches@redhat.com>
2026-03-27 13:25:19 -07:00
Kyle SayersandGitHub 648edcf729 [QeRL] Compose online quantization with quantized reloading (#38032)
Signed-off-by: Kyle Sayers <kylesayrs@gmail.com>
2026-03-27 13:22:33 -07:00
7ba425e916 Add short flag -sc for --speculative-config argument (#38380)
Co-authored-by: Claude <noreply@anthropic.com>
2026-03-27 12:04:22 -07:00
Gregory ShtrasbergandGitHub b8665383df [ROCm] Fix GPT-OSS import for triton 3.6 (#37453)
Signed-off-by: Gregory Shtrasberg <Gregory.Shtrasberg@amd.com>
2026-03-27 18:00:57 +00:00
0e9358c11d {ROCm]: gpt-oss fusion/padding fixes (#38043)
Signed-off-by: Andreas Karatzas <akaratza@amd.com>
Signed-off-by: Rohan138 <rohanpotdar138@gmail.com>
Signed-off-by: Rohan Potdar <66227218+Rohan138@users.noreply.github.com>
Co-authored-by: Andreas Karatzas <akaratza@amd.com>
2026-03-27 12:19:15 -04:00
Harry MellorGitHubgemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
21d2b53f88 Remove need for explicit \n in docstring lists for --help formatting (#38350)
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2026-03-27 08:38:00 -07:00
Jonas M. KüblerandGitHub 98e7f223b9 enable skipping of SW attention layers when using FP8 KV cache (#33695)
Signed-off-by: Jonas Kuebler <kuebj@amazon.com>
2026-03-27 07:25:02 -06:00
b111f8a61f fix(security): Add VLLM_MAX_N_SEQUENCES environment variable and enforce limit (#37952)
Signed-off-by: jperezde <jperezde@redhat.com>
Signed-off-by: Russell Bryant <rbryant@redhat.com>
Co-authored-by: Russell Bryant <rbryant@redhat.com>
2026-03-27 09:02:10 -04:00
Sage MooreandGitHub 497e234d38 [EPLB] Cleanup the transfer logic for the various eplb maps (#34520)
Signed-off-by: Sage Moore <sagmoore@redhat.com>
Signed-off-by: Sage Moore <sage@neuralmagic.com>
2026-03-27 10:18:46 +01:00
dtcandGitHub 6287e7fa20 [P/D] Mooncake: Add unit tests and minor fixes for mooncake connector (#36946)
Signed-off-by: Tianchen Ding <dtcccc@linux.alibaba.com>
2026-03-27 09:26:40 +01:00
Shengqi ChenGitHubClaude Opus 4.6gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>Copilot
84e439a9cb [CI/Build] Move nightly wheel index generation to a single post-build step (#38322)
Signed-off-by: Shengqi Chen <harry-chen@outlook.com>
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
2026-03-27 07:44:18 +00:00
Yuichiro UtsumiGitHubgemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
a1746ff9ec [Doc] Clarify Helm chart location in deployment guide (#38328)
Signed-off-by: Yuichiro Utsumi <utsumi.yuichiro@fujitsu.com>
Signed-off-by: Yuichiro Utsumi <81412151+utsumi-fj@users.noreply.github.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
2026-03-27 15:43:02 +08:00
Flora FengandGitHub aee4c14689 [Bugfix] Fix Hermes tool parser when stream interval > 1 (#38168)
Signed-off-by: sfeng33 <4florafeng@gmail.com>
2026-03-27 14:42:26 +08:00
Bowen BaoandGitHub 0ae89f18fd [Refactor] Move FusedMoE hidden_size roundup to quant_method (#34285)
Signed-off-by: Bowen Bao <bowenbao@amd.com>
2026-03-26 23:38:26 -07:00
wenjun liuandGitHub c2b17d71af [CI] Add xpu auto-label rule for Intel GPU/XPU PRs (#38320)
Signed-off-by: wendyliu235 <wenjun.liu@intel.com>
2026-03-27 14:22:38 +08:00
Li, JiangandGitHub becaed6ec8 [CPU] Support CT W4A16 on CPU MP kernel (#38219)
Signed-off-by: jiang1.li <jiang1.li@intel.com>
2026-03-27 14:15:28 +08:00
Xiaoshuang WangandGitHub a8eab8f30d [Model] Extract GatedDeltaNetAttention into shared layer for Qwen3Next and Qwen3.5 (#37975)
Signed-off-by: wxsIcey <1790571317@qq.com>
Signed-off-by: Icey <1790571317@qq.com>
2026-03-27 14:13:21 +08:00
cjackalandGitHub 2babac0bed [frontend] dump openai responses type by alias (#38262)
Signed-off-by: cjackal <44624812+cjackal@users.noreply.github.com>
2026-03-27 05:58:20 +00:00
Or OzeriandGitHub 7cc302dd87 [kv_offload+HMA][7/N]: Support register_kv_caches for hybrid models (#37853)
Signed-off-by: Or Ozeri <oro@il.ibm.com>
2026-03-27 08:38:33 +03:00
Woosuk Kwon a17a1f12dc inlined dsv3.2
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
2026-03-27 05:20:23 +00:00
999dfc1622 [Bugfix] Offload blocking tokenizer ops to shared thread pool to unblock event loop (#34789)
Signed-off-by: Bvicii <yizhanhuang2002@gmail.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
2026-03-26 22:17:00 -07:00
wenjun liuandGitHub d86060122a [CI/Build] enable Intel XPU test flow with prebuilt image (#37447)
Signed-off-by: wendyliu235 <wenjun.liu@intel.com>
2026-03-26 18:16:04 -07:00