 Liangsheng Yin
|
ec6f8d61f7
|
[Spec] Async-assert probes across EAGLE/MTP; zero tgt_cache_loc (#26335)
|
2026-05-26 11:34:03 -07:00 |
|
|
Venkatesh GuduruandvguduruTT
|
6c8128650e
|
[Bugfix] Fix flashinfer_cutlass MoE crash when intermediate_size_per_partition is not 16-aligned (#22627)
Co-authored-by: vguduruTT <venkatesh.guduru@mulitcorewareinc.com>
|
2026-05-26 16:58:33 +00:00 |
|
|
Zhangheng
|
38f32c38ab
|
[UnifiedRadixTree]: Support L3 HiStorage framework (#26062)
|
2026-05-26 22:38:00 +08:00 |
|
|
Jun Liu
|
c47f0e7cdd
|
[PD] Fix top logprobs crash in prefill path (#26299)
|
2026-05-26 22:01:10 +08:00 |
|
|
Shangming Cai
|
c8c1aed5e9
|
[PD] Fix cross-rank queue divergence by gating metadata readiness before all-reduce (#26394)
Signed-off-by: Shangming Cai <csmthu@gmail.com>
|
2026-05-26 21:59:42 +08:00 |
|
|
Yongfei Xu
|
98eb84497d
|
[PP] Skip PP output communication for pure chunked prefill batches (#26148)
|
2026-05-26 21:59:18 +08:00 |
|
|
Michael
|
9409969fd5
|
Revert "[perf][spec decoding] Skip full-vocab softmax in EAGLE draft when topk == 1 (#26235)" (#26358)
|
2026-05-26 02:47:52 -07:00 |
|
|
fzyzcjy
|
d9c82934c8
|
Extract Scheduler init methods and add skills to enforce the splitting requirements (#26271)
|
2026-05-26 17:45:09 +08:00 |
|
|
Chao Shi
|
48f3264807
|
[HiCache]: Check return code of cudaHostRegister (#26301)
|
2026-05-26 17:44:15 +08:00 |
|
|
roikoren755
|
e958f4561f
|
[feat] Support extra_buffer in Mamba2-based models (#15829)
Signed-off-by: Roi Koren <roik@nvidia.com>
|
2026-05-26 16:03:29 +08:00 |
|
|
Liangsheng Yin
|
7e6e5efe51
|
Revert "fix(tool_call): normalize non-standard JSON Schema types in tool params" (#26379)
|
2026-05-26 00:49:41 -07:00 |
|
|
Zhonghua Deng
|
dabdd91ef3
|
[EPD] Cross-request batching for image/audio encoder (#25964)
|
2026-05-26 15:38:59 +08:00 |
|
|
Yuan Luoandluoyuan.luo
|
d34d4d9f5f
|
[GDN] Support SM100 CuTeDSL GDN Prefill Kernel (#26200)
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
|
2026-05-26 15:38:29 +08:00 |
|
|
Xinyuan Tong
|
64c7c6851b
|
fix(tool_call): normalize non-standard JSON Schema types in tool params (#23476)
|
2026-05-26 15:23:59 +08:00 |
|
|
chengchao23
|
137168539a
|
[Perf][Moe]improve cutlass_moe_fp4 performance by using apply_router_weight_on_i… (#19493)
|
2026-05-26 00:07:49 -07:00 |
|
|
+2
|
3f5e2c7688
|
[AMD] Dsv4/pr2 compressor opt (#26208)
Co-authored-by: wunhuang <wunhuang@amd.com>
Co-authored-by: Thomas Wang <1am9trash@gmail.com>
Co-authored-by: Xinyi Song <86638975+RolaoDenthu@users.noreply.github.com>
Co-authored-by: HaiShaw <hixiao@gmail.com>
Co-authored-by: amd-danli103 <danli103@amd.com>
Co-authored-by: Lin, Soga <soga.lin@amd.com>
Co-authored-by: Raiden-Makoto <Raiden-Makoto@users.noreply.github.com>
Co-authored-by: Hubert Lu <55214931+hubertlu-tw@users.noreply.github.com>
Co-authored-by: yichiche@amd.com <jacky.cheng>
Co-authored-by: yctseng0211 <yctseng@amd.com>
Co-authored-by: Bingxu Chen <bingxche@amd.com>
|
2026-05-25 23:54:40 -07:00 |
|
|
Shaojun Zhou
|
7c0fbc8c2e
|
fix: fix fa3 cross-attention batched-decode for per-request varlen encoder (#25045)
|
2026-05-26 14:26:15 +08:00 |
|
|
Mick
|
63a89bf1a9
|
[kernel] reuse wna16 marlin moe workspace (#26112)
|
2026-05-26 13:19:13 +08:00 |
|
|
Yuxuan Zhang
|
7ef06bfc06
|
GLM-4.7-Flash: standalone MLA impl and MLA NextN/MTP (#26088)
|
2026-05-26 13:17:39 +08:00 |
|
|
xutizhou
|
59cad671e2
|
Support DeepSeek V4 DeepEP Waterfill (#25391)
|
2026-05-25 21:04:26 -07:00 |
|
|
mispa-msandMick
|
3142278c5f
|
[diffusion] feat: layerwise NVTX markers for Nsight Systems profiling (#25683)
Co-authored-by: Mick <mickjagger19@icloud.com>
|
2026-05-26 11:24:27 +08:00 |
|
|
Chetan Kumar Verma
|
156d1af23a
|
[Intel GPU] Fix incorrect KV-cache page table for local attention when page_size > 1 (#23757)
|
2026-05-26 11:02:21 +08:00 |
|
|
Mick
|
8f2a4e70f8
|
[SRT] minor: reuse req input id array for unpadded ids (#26232)
|
2026-05-26 08:58:32 +08:00 |
|
|
Liangsheng Yin
|
8805f4cf16
|
Fail-fast on PD subprocess exit and scheduler exception (#26298)
|
2026-05-25 16:42:50 -07:00 |
|
|
Ke Bao
|
e7b12fe6fa
|
Fix stale forward_metadata leak in DP attn unpadded idle batch (#26313)
|
2026-05-25 16:04:00 -07:00 |
|
|
Ziang Li
|
2b9dd9c8b3
|
[FlashInfer v0.6.10] [RL] [DSv32] [GLM-5] Add --dsa-topk-backend and integrate FlashInfer and pytorch topk (#22851)
|
2026-05-25 13:08:03 -07:00 |
|
|
Ke Bao
|
b13d3d18c6
|
Refactor HiCache stack dispatch into strategies (#26295)
|
2026-05-26 00:06:17 +08:00 |
|
|
Shangming Cai
|
2aa6995308
|
[CI] Enable EPD CI for EPD architecture enhancements (#26281)
|
2026-05-25 23:52:58 +08:00 |
|
|
Xiaoyu ZhangandBBuf
|
121cc09405
|
[diffusion] Add CFG gating for denoising (#25848)
Co-authored-by: BBuf <bbuf@example.com>
|
2026-05-25 22:57:09 +08:00 |
|
|
Xiaoyu ZhangandBBuf
|
85f9522e36
|
[diffusion] Cache fp32 layernorm params (#25847)
Co-authored-by: BBuf <bbuf@example.com>
|
2026-05-25 22:56:39 +08:00 |
|
|
Makcum888eandronnie_zheng
|
0801cc05ed
|
[Diffusion][NPU] Disaggregation diffusion stages support for NPU (#25895)
Co-authored-by: ronnie_zheng <zl19940307@163.com>
|
2026-05-25 13:51:25 +03:00 |
|
|
Liangsheng Yin
|
3e67398a96
|
Zero req_pool_indices padding in cuda-graph populate (#26292)
|
2026-05-25 03:29:05 -07:00 |
|
|
Xiaoyu Zhang
|
533ef41112
|
[Diffusion] Default NVFP4 backend to FlashInfer TRTLLM (#25523)
|
2026-05-25 18:14:06 +08:00 |
|
|
Mick
|
c05756da7a
|
[SRT] fix flashInfer allreduce fusion not used on blackwell (#26197)
|
2026-05-25 18:07:15 +08:00 |
|
|
Qiaolin Yu
|
a77449f86d
|
[perf][spec decoding] Skip full-vocab softmax in EAGLE draft when topk == 1 (#26235)
|
2026-05-25 02:06:48 -07:00 |
|
|
Shangming Cai
|
7f2829af39
|
chore: bump mooncake version to 0.3.11.post1 (#25989)
|
2026-05-25 16:51:03 +08:00 |
|
|
xdtbynd
|
0942011665
|
[NPU] Add torchaudio dependency for NPU platform (#26267)
|
2026-05-25 16:30:12 +08:00 |
|
|
Yuhui Liang
|
ca029e816b
|
Fix missing idle-batch handling in prepare_mlp_sync_batch_raw (#25404)
|
2026-05-24 23:53:54 -07:00 |
|
|
Erik Wijmans
|
87e69d57c4
|
[lora] Fix overlap loading for cancelled requests (#25413)
|
2026-05-25 15:18:36 +09:00 |
|
|
Xia WeiwenandMa Mingfei
|
2bd3ac0b5d
|
[XPU] fix correctness issue of GDN triton kernel for XPU (#26065)
Co-authored-by: Ma Mingfei <mingfei.ma@intel.com>
|
2026-05-25 13:18:11 +08:00 |
|
|
Qiaolin Yu
|
ec6fcb93cb
|
[perf][spec decoding] Skip common_template in TRTLLMMLAMultiStepDraftBackend init (#26241)
|
2026-05-24 21:36:16 -07:00 |
|
|
Ethan ZHUandZhangheng
|
e86fdf3a3c
|
[Bug Fix][HiCache] TreeNode.get_prefix_hash_values @lru_cache can return mutated list (#26177)
Co-authored-by: Zhangheng <hzh0425@apache.org>
|
2026-05-25 11:15:10 +08:00 |
|
|
Ma Mingfei
|
821d5f4a5b
|
[CPU] add faster KV-cache writes (#25874)
|
2026-05-25 10:28:52 +08:00 |
|
|
Mick
|
e1463bb2c2
|
[VLM] try to reuse precomputed padded input ids in scheduler instead of padding (#26097)
|
2026-05-25 10:27:04 +08:00 |
|
|
Liangsheng Yin
|
850887dc63
|
[Spec] fix EAGLE v2 verify metadata init order on non-cuda-graph path (#26244)
|
2026-05-24 18:49:33 -07:00 |
|
|
Mick
|
64e2b54a8f
|
[VLM] feat: accept grid_thws from preprocessed metadata for kimi (#26149)
|
2026-05-25 09:10:25 +08:00 |
|
|
Mick
|
72c1582d4e
|
[VLM] fix: fix only the grids from last split mm item is collected for qwen-vl (#26094)
|
2026-05-25 09:09:46 +08:00 |
|
|
Liangsheng Yin
|
ed179bf9b2
|
[dsv4] fix multi-step draft on non-cuda-graph path (#26239)
|
2026-05-24 17:04:18 -07:00 |
|
|
Ming Yang
|
85471d253d
|
Add --disable-attn-tp-gather opt-out for model-managed SP (#26047)
|
2026-05-24 15:53:51 -07:00 |
|
|
Lianmin Zheng
|
93fa577bb9
|
Clean up server startup log noise (#26205)
|
2026-05-24 14:35:15 -07:00 |
|