Commit Graph

8017 Commits

Author SHA1 Message Date
Sam (Kesen Li)
73e93bebd6 [1/4] NVFP4 KV cache: quantization strategy abstraction and kernel (#21954) 2026-04-29 01:45:48 -07:00
danielafrimi
8327270c72 [Kernel] Support FlashInfer TRTLLM-Gen fused MoE for non-gated FP4 & FP8 (Nemotron) (#21321) 2026-04-29 01:28:21 -07:00
Liangsheng Yin
bd448e51bd [Spec] Split accept_length into num_accepted_drafts and num_accepted_tokens (#23962) 2026-04-29 00:02:22 -07:00
shuwenn
2c41ef4c93 [HiCache] feat: add draft KV cache backing for L2/L3 (#21125) 2026-04-28 23:47:31 -07:00
Mick
71d2227a78 [diffusion] CI: update ground truth with official output (#23714) 2026-04-29 13:51:31 +08:00
JieTang66
5a94088178 Fix typo: page_first_kv_spilt -> page_first_kv_split (#23983)
Co-authored-by: JieTang <tangjie66@huawei.com>
2026-04-29 12:08:05 +08:00
ZeyuanChen2000
23cfa9e4c8 [NPU] fix rope_theta get error for baichuan2-13b-chat model (#21543) 2026-04-29 12:01:03 +08:00
Yuhao Yang
b437f6be48 model: Nemotron-omni-v3-alias (#23857) 2026-04-29 11:08:23 +08:00
Qiaolin Yu
f57ec8d6ef [spec decoding] add extra attribute 'spec_hidden_size' (#23890) 2026-04-28 19:54:50 -07:00
Lianmin Zheng
2a771a40ac Add engine_type label to tokenizer manager metrics (#23978) 2026-04-28 19:52:58 -07:00
Rahul Vijayaraghavan
1aee04e7df [XPU] Support apply_router_weight_on_input for Llama4 for fused_experts (#22654)
merge this one as it is xpu only change.
2026-04-29 10:44:48 +08:00
ccullen-cert
1e14bd6f36 Fix for CVE-2026-5760 (#23660) 2026-04-28 19:39:44 -07:00
Hubert Lu
e5da200d0a [AMD] Fix Aiter RMSNorm layout handling (#23974) 2026-04-28 19:28:46 -07:00
Chandrakant Khandelwal
0ac23cffac Add intel_xpu as backend for GptOssForCausalLM, enabled for bf16 models with torch native backend (#12771)
Co-authored-by: Ma Mingfei <mingfei.ma@intel.com>
2026-04-29 10:21:52 +08:00
iridiumine
08699bb1b2 [NPU] Fix DeepEP LL dispatch BF16 flag and skip triton kernel on NPU for Qwen3.5 (#23815)
Co-authored-by: iridiumine <iridiumine@users.noreply.github.com>
2026-04-29 10:19:24 +08:00
Xiaoyu Zhang
13afe8acdf [codex] Enable Qwen3-Next MoE all-reduce fusion (#23619) 2026-04-29 09:11:35 +08:00
Lianmin Zheng
14b4e6fa69 Support --model as alias for --model-path in CLI (#23894) 2026-04-28 17:23:27 -07:00
shuwenn
233048212a [HiCache][SPEC] fix: normalize storage prefetch key (#23631) 2026-04-28 15:53:20 -07:00
Khoa Pham
ddcacaf1bd Fix failing test_nvidia_nemotron_3_nano by fixing test_grouped_topk (#23874) 2026-04-28 15:03:58 -07:00
Liangsheng Yin
cf0061da43 [Spec] Fix spec_accept_rate and unify accept/draft naming (#23530) 2026-04-28 14:40:04 -07:00
shuwenn
3e1c5e1b74 [HiCache][SPEC] fix: empty key after page alignment in match_prefix (#23387)
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-04-28 14:06:55 -07:00
Lewis
9814cc89ce [Fix] NVFP4 qwen3.5 quant error fix by add packed_modules_mapping (#23471)
Co-authored-by: 百麒 <yaozhong.lyz@alibaba-inc.com>
2026-04-28 13:36:09 -07:00
robellliu-dev
914ef7c7f3 Fix multimodal /v1/embeddings Jinja chat template handling (#20835)
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
Co-authored-by: Xinyuan Tong <115166877+JustinTong0323@users.noreply.github.com>
2026-04-28 13:05:45 -07:00
Qingfu Wen
dc1eac4903 [MUSA][Diffusion] Fix fa3 API on MT MUSA (#23646) 2026-04-28 13:01:35 -07:00
AlonKejzman
66ea0aee7f tokenizer: Add fastokens support (#23753) 2026-04-28 11:43:10 -07:00
Mick
144038fbae [diffusion] chore: change default seed to 42 (#23836) 2026-04-28 20:39:23 +08:00
Xiaoyu Zhang
7824903417 [SKILL] Sync SGLang skill docs (#23921) 2026-04-28 17:05:36 +08:00
Yinzuo Jiang
71160e4ddb feat(observability): add OpenTelemetry tracing for pipeline parallelism (#23169)
Signed-off-by: Yinzuo Jiang <jiangyinzuo@foxmail.com>
2026-04-28 17:05:23 +08:00
Xun Sun
9a53ab3d6d [6/N] (Elastic EP) Recover failed ranks (#15771) 2026-04-28 00:44:26 -07:00
yaya159456
b8a2dcd300 fix: resolve tensor file overwrite between target and draft models (#21694)
Co-authored-by: jiangguangya <jiangguangya@baidu.com>
Co-authored-by: Khoa Pham <khoa.pham@radixark.ai>
2026-04-27 23:40:49 -07:00
Xiaoyu Zhang
6fbad22feb Remove smoke wording from tests and comments (#23355) 2026-04-28 12:05:27 +08:00
JoyFuture
1a55646dcd [Feature] Xiaomi MiMo-V2.5-Pro day0 support (#23808) 2026-04-28 11:43:29 +08:00
Baizhou Zhang
c1d1412333 [Flashinfer] Integrate flashinfer router gemm for sm103 (#23285) 2026-04-27 20:37:56 -07:00
chenkaiyue
3066ba8167 fix(hicache): add retry logic for MooncakeStore warmup (#17195)
Co-authored-by: Teng Ma <sima.mt@alibaba-inc.com>
Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu>
2026-04-27 18:50:43 -07:00
cen121212
b1e1fe8eee 【NPU】【bugfix】accuracy fix when enable both nsa cp and prefixcache (#23268) 2026-04-28 09:08:28 +08:00
看海的人
9ffc0cc67e [NPU] Support GLM-4.5V (#22961) 2026-04-28 09:08:19 +08:00
PiteXChen
4a04a9818e 【hicache】Optimize HiCache prefetch logic: adapt to remaining available memory when memory is insufficient (#16370)
Signed-off-by: CLFutureX <chenyongqyl@163.com>
Co-authored-by: Zhiqiang Xie <xiezhq@stanford.edu>
2026-04-27 17:18:35 -07:00
Byron Hsu
27659ea8c8 [PD+Pause] Remove redundant post processing (#23886)
Co-authored-by: root <root@slurm-h200-206-011.slurm-compute.tenant-slurm.svc.cluster.local>
2026-04-27 17:01:26 -07:00
Byron Hsu
cb0429f253 [Disagg] Finalize routed_experts_output in process_batch_result_disagg_prefill (#23885)
Co-authored-by: Byron Hsu <byron@periodiclabs.ai>
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-27 16:40:05 -07:00
Bi Xue
41181b6238 [sgl] copy mm_input in piecewise cuda graph when eagle3 is on (#23613) 2026-04-27 13:35:19 -07:00
Kurt Shuster
f34c20af86 [VLM] Fix Kimi-K2.5 CPU path: rename grid_thws -> image_grid_thw (#23501)
Co-authored-by: Ethan (Yusheng) Su <yushengsu.thu@gmail.com>
2026-04-27 20:34:28 +00:00
Vladislav Nosivskoy
28ee08c172 [HiCache] Add synchronization for context parallelism (#20460)
Signed-off-by: Vladislav Nosivskoy <vladnosiv@gmail.com>
2026-04-28 02:13:07 +08:00
Jonah Bernard
4fc5ebf0b7 [Chore] Remove deadcode in prefill delayer (#23389)
Co-authored-by: Jonah Bernard <96398205+Jonahcb@users.noreply.github.com>
2026-04-27 09:59:36 -07:00
ranjiewen
f2b84b90ac [npu]fix: qwen3-next w8a8 precision bugs (#21698) 2026-04-27 18:14:33 +08:00
Lianmin Zheng
8536d4b402 Clean up noisy startup warnings from third-party deps (#23669) 2026-04-27 03:10:46 -07:00
Shenxiu Liu
a3fc982ba7 [Whisper] Automatic language detection via structured generation (#22997)
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
2026-04-27 15:54:41 +08:00
Xiaoyu Zhang
5f47cae1a0 add H100 configs for GLM-4.7-Flash (#23719) 2026-04-27 15:07:39 +08:00
Colin Z
d49561b8ae [AMD] Fix Kimi-K2.6 Quark MXFP4 loading prefix and packed module mapping (#23408) 2026-04-26 23:56:15 -07:00
Praneth Paruchuri
b7113cadb1 [Bug Fix] Reject pp_max_micro_batch_size=0 to prevent silent deadlock on generate() (#23799) 2026-04-27 13:36:04 +08:00
Xinyuan Tong
e5198386bd Upgrade transformers from 5.5.4 to 5.6.0 (#23525) 2026-04-26 22:33:54 -07:00