|
f4e7a98fe5
|
[HiCache] feat: support draft offload for mooncake (#24984)
Co-authored-by: huangtingwei9988 <141888744+huangtingwei9988@users.noreply.github.com>
Co-authored-by: stmatengss <11641725+stmatengss@users.noreply.github.com>
|
2026-06-02 21:42:04 -07:00 |
|
|
CrazyCoder
|
c3aaafc5f2
|
[Bugfix] Clean up failed NIXL sender state (#27011)
|
2026-06-03 12:15:15 +08:00 |
|
|
zhaoshangandShangming Cai
|
3e681d7fff
|
Add per-rank staggered weight loading for improved TP I/O concurrency (#26937)
Signed-off-by: zhaoshang <zhaoshangsjtu@linux.alibaba.com>
Co-authored-by: Shangming Cai <csmthu@gmail.com>
|
2026-06-03 11:25:21 +08:00 |
|
|
zijiexiaandYihao Wang
|
1ebc7438ac
|
model: support Command A plus (#26106)
Co-authored-by: Yihao Wang <42559837+AgainstEntropy@users.noreply.github.com>
|
2026-06-03 11:23:04 +08:00 |
|
|
Mick
|
71a747cf15
|
[diffusion] fix: fix lingbot realtime consistency gt pin (#27080)
|
2026-06-03 10:29:37 +08:00 |
|
|
Chi McIsaac
|
3d54056391
|
[diffusion] optimize: optimize Cosmos3 i2v latent prep (#27084)
|
2026-06-03 10:23:37 +08:00 |
|
|
Mick
|
3715a07e87
|
[diffusion] improve: clamp wanvae decode output in place (#27086)
|
2026-06-03 10:16:10 +08:00 |
|
|
Mick
|
f0e18be0ba
|
[diffusion] improve: use conv2d width padding in wanvae (#27081)
|
2026-06-03 10:14:50 +08:00 |
|
|
Brayden Zhongandb8zhong
|
13852d3f31
|
Support NextN = 2/4 in DSV32 (#24870)
Co-authored-by: b8zhong <b8zhong@users.noreply.github.com>
|
2026-06-02 19:06:06 -07:00 |
|
|
MingxuZh
|
b678448b8a
|
ci(xeon): merge 2 partitions into 1 job to reduce runner contention (#26904)
|
2026-06-03 09:46:27 +08:00 |
|
|
gaopengff
|
eda21f6839
|
Add fused_rope and for xpu (#25773)
|
2026-06-03 09:41:42 +08:00 |
|
|
Mick
|
83bc776612
|
[diffusion] optimize: preserve dtype in wanvae nearest upsample (#27077)
|
2026-06-03 08:32:28 +08:00 |
|
|
Qiaolin Yu
|
c55548ba11
|
[perf] Replicate embed_tokens to drop the post-embed all-reduce (#26970)
|
2026-06-02 16:48:18 -07:00 |
|
|
Alison ShaoandCheng Wan
|
76c9899da7
|
Fix hybrid linear attention misrouting plain-RadixAttention linear layers to the full backend (Ring-2.5-1T) (#26623)
Co-authored-by: Cheng Wan <54331508+ch-wan@users.noreply.github.com>
|
2026-06-02 16:24:49 -07:00 |
|
|
Hubert Lu
|
72929c7000
|
[AMD] Enable AITER custom all-gather on ROCm (#25093)
|
2026-06-02 15:57:37 -07:00 |
|
|
Khoa Pham
|
a711c57a32
|
[Spec] Fix Gemma 4 MTP with trtllm_mha crash issue (#26966)
|
2026-06-02 14:37:51 -07:00 |
|
|
Alison Shao
|
365cc2ade5
|
jit_kernel tests: bump multiprocess_test timeout 90s -> 240s (cold JIT cache) (#26994)
|
2026-06-02 17:33:52 -04:00 |
|
|
Baizhou Zhang
|
68caf49154
|
Update sgl-deep-gemm to 0.1.2 (#26993)
|
2026-06-02 13:51:54 -07:00 |
|
|
Hanming Lu
|
b603f08c0c
|
[DP] Fix FlashInfer dispatcher workspace sizing and set_dp_buffer_len (#26643)
|
2026-06-02 13:25:20 -07:00 |
|
|
Bi Xue
|
9e717cae46
|
[sglang] Fix Mamba COW over-releasing SWA locks (cascade-evict assert crash) (#27038)
|
2026-06-03 01:42:04 +08:00 |
|
|
Muqi Li
|
6ba31e33e6
|
feat(api): add require_reasoning field for engine's generate api (#27019)
|
2026-06-02 17:36:29 +00:00 |
|
|
Cheng WanandClaude Opus 4.7
|
99da43b900
|
[refactor] init_forward_metadata 3-method ABC + side-channel removal + ForwardMetadata type rename (#26735)
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
|
2026-06-02 10:33:33 -07:00 |
|
|
Ilia Yastrebov
|
6c69756fa8
|
NIXL: use prep+make API to improve performance (#26406)
|
2026-06-02 18:36:35 +02:00 |
|
|
Mick
|
22043b917b
|
[diffusion] CI: ad lingbot case (#27055)
|
2026-06-02 22:26:51 +08:00 |
|
|
Ke Bao
|
38d4c9ba88
|
Improve type annotations in unified radix cache (#26948)
|
2026-06-02 21:17:50 +08:00 |
|
|
Mick
|
64a1dec8b6
|
[diffusion] feat: add realtime webui super resolution controls (#27026)
|
2026-06-02 20:29:21 +08:00 |
|
|
Mick
|
ce7da7397a
|
[diffusion] optimize: optimize cosmos3 (#27041)
|
2026-06-02 19:47:42 +08:00 |
|
|
Librua
|
c2eea4d7b3
|
[Bugfix] Fix orphaned aborted prefill bootstrap requests in PP disaggregation (#27028)
|
2026-06-02 19:03:46 +08:00 |
|
|
Mick
|
3394931044
|
[diffusion] optimize: optimize lingbot performance (#27023)
|
2026-06-02 18:33:06 +08:00 |
|
|
Mick
|
a777672939
|
[diffusion] feat: enable parallel decode for cosmos3(#27037)
|
2026-06-02 18:18:18 +08:00 |
|
|
|
84e1108312
|
Optimize ngram decode id computation (#24757)
Co-authored-by: Codex <codex@example.com>
Co-authored-by: BBuf <xiaoyu.zhang@radixark.net>
|
2026-06-02 17:37:34 +08:00 |
|
|
pengduriceandgithub-actions[bot]
|
f651b48764
|
Apply apply_group_norm_silu to LTX-2 latent upsampler (#26045)
Signed-off-by: pengdurice <pengduhit@gmail.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
|
2026-06-02 17:28:57 +08:00 |
|
|
Xiaoyu ZhangandBBuf
|
3ea1ba5b15
|
[GDN] Optimize prefill QKV split dispatch (#26206)
Co-authored-by: BBuf <xiaoyu.zhang@radixark.net>
|
2026-06-02 16:48:31 +08:00 |
|
|
Xiaoyu ZhangandBBuf
|
559581b383
|
[codex] Centralize Triton utility kernels (#26000)
Co-authored-by: BBuf <xiaoyu.zhang@radixark.net>
|
2026-06-02 16:47:45 +08:00 |
|
|
Bruce Changlong XuandKe Bao
|
172bd8e6b9
|
[scheduler] Zero gen_throughput and flush KV events on pause (#24003)
Co-authored-by: Ke Bao <ispobaoke@gmail.com>
|
2026-06-02 16:43:04 +08:00 |
|
|
cctryandgemini-code-assist[bot]
|
b55570d38e
|
[PD] Optimistic prefill (#26780)
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
|
2026-06-02 01:16:14 -07:00 |
|
|
Charles Chen
|
5ae8d286d2
|
perf(gemma4): single-launch fused router (topk + softmax + scale) (#26502)
|
2026-06-02 16:00:17 +08:00 |
|
|
|
3e993f6140
|
[PD]: Support HiCache prefetching and pd-incremental transfer on decode side (#26227)
Co-authored-by: huangtingwei <141888744+huangtingwei9988@users.noreply.github.com>
Co-authored-by: Shangming Cai <csmthu@gmail.com>
Co-authored-by: 晟海 <huangtingwei.htw@antgroup.com>
|
2026-06-02 15:40:10 +08:00 |
|
|
Jinyan ChenandJinyan Chen
|
301bcf0872
|
Add FP4 Indexer for DeepSeek V4 (#26209)
Co-authored-by: Jinyan Chen <jinyanc@nvidia.com>
|
2026-06-02 00:14:38 -07:00 |
|
|
Mick
|
1033d835ff
|
[diffusion] optimize: reduce cosmos3 denoise overhead (#26973)
|
2026-06-02 14:23:02 +08:00 |
|
|
Mick
|
3b26644bc4
|
[diffusion] misc: add realtime-webui (#26959)
|
2026-06-02 14:13:02 +08:00 |
|
|
Mick
|
2fc548f250
|
[diffusion] model: support lingot-world (#26954)
|
2026-06-02 13:52:49 +08:00 |
|
|
Liangsheng Yin
|
f531bd7ff3
|
[Bug] Fix circular import in forward_batch_info from runtime cp_utils import (#27014)
|
2026-06-01 22:51:32 -07:00 |
|
|
Thomas Wang
|
d15a2dc72c
|
[AMD] dpsk-v4 swa loc cache support (#26931)
|
2026-06-01 22:37:07 -07:00 |
|
|
|
4226a6f13a
|
[AMD] Fix GPT-OSS MXFP4 accuracy on ROCm AITER path (#26884)
Co-authored-by: wunhuang <wunhuang@amd.com>
Co-authored-by: Bingxu Chen <bingxche@amd.com>
|
2026-06-01 22:30:43 -07:00 |
|
|
Khoa PhamandCursor
|
08526c7fca
|
[Spec] FrozenKVMTP fold assistant seed into captured draft graph (#25539)
Co-authored-by: Cursor <cursoragent@cursor.com>
|
2026-06-01 22:27:25 -07:00 |
|
|
Ethan ZHUandZhangheng
|
594ec6335d
|
[Bug Fix][HiCache] Drop @lru_cache on UnifiedTreeNode.get_prefix_hash_values (#26939)
Co-authored-by: Zhangheng <hzh0425@apache.org>
|
2026-06-02 12:38:28 +08:00 |
|
|
popsiclexuandpopsiclexu
|
951fa05a09
|
[MoE] Support BF16 standard A2A with DeepGEMM runner (#26473)
Co-authored-by: popsiclexu <zhenxue.xu@mthreads.com>
|
2026-06-01 20:40:38 -07:00 |
|
|
Teng MaandZijie Xia
|
b562da0d9f
|
[PD] docs: clarify disaggregation IB device formats (#25521)
Co-authored-by: Zijie Xia <zijie.xia@radixark.ai>
|
2026-06-02 11:38:33 +08:00 |
|
|
ybyang
|
9fe8b72912
|
Speed up DeepGEMM JIT warmup with per-PP-rank parallel compile (#26567)
|
2026-06-01 19:51:27 -07:00 |
|