yudian0504 3d312643b9 [BUGFIX] Fix CP residual size mismatch crash when tp_size == attn_cp_size (#21170)

2026-03-23 00:12:58 -07:00

..

batch_invariant_ops

[Auto Sync] Update batch_invariant_ops.py (20260221) (#19098)

2026-02-20 18:27:31 -08:00

batch_overlap

[AMD][TBO] Fix mori ep dual stream accuracy (#19888)

2026-03-18 02:00:55 -07:00

checkpoint_engine

update pre-commit config (#18860)

2026-02-16 00:18:31 +08:00

compilation

Simplify server startup output (#20885)

2026-03-19 13:11:37 -07:00

configs

Add Mistral Small 4 (Pixtral) support (#20708)

2026-03-18 14:15:32 -07:00

connector

Remove sync points in mamba cache + prefill cudagraph plumbing for DP (#19639)

2026-03-02 15:03:42 +08:00

constrained

fix: propagate grammar errors and improve llguidance backend (#20467)

2026-03-15 16:11:18 -07:00

debug_utils

Add e2e demo test in dump comparator (#21031)

2026-03-20 22:41:01 +08:00

disaggregation

Add logging for BootstrapServer for CI diagnosis (#20844)

2026-03-19 14:42:12 +08:00

distributed

Support Qwen3 MoE context parallel (#18233)

2026-03-22 01:27:20 -07:00

dllm

Decouple scheduler log printing from metrics collection (#20107)

2026-03-07 22:09:10 -08:00

elastic_ep

[5/N] (Elastic EP) Use GPU P2P to exchange expert weights during EPLB as much as possible (#12068)

2026-03-16 18:40:58 +08:00

entrypoints

Fix streaming logprobs corruption caused by shared mutable list reference (#21030)

2026-03-21 00:18:48 -07:00

eplb

[5/N] (Elastic EP) Use GPU P2P to exchange expert weights during EPLB as much as possible (#12068)

2026-03-16 18:40:58 +08:00

function_call

[feat] Enhance Kimi-K2/K2.5 function call and reasoning detection (#19552)

2026-03-19 12:57:57 -07:00

grpc

[gRPC] Extract gRPC servicer into standalone package (#20478)

2026-03-13 09:13:29 -07:00

hardware_backend/npu

[NPU] add new fusion operator DispatchFFNCombine (#20245)

2026-03-18 15:22:04 +08:00

layers

[BUGFIX] Fix CP residual size mismatch crash when tp_size == attn_cp_size (#21170)

2026-03-23 00:12:58 -07:00

lora

[LoRA][II] Add fused MOE LoRA Triton kernel and tests (#19711)

2026-03-18 19:58:14 -07:00

managers

HiSparse for Sparse Attention (#20343)

2026-03-22 23:09:31 -07:00

mem_cache

HiSparse for Sparse Attention (#20343)

2026-03-22 23:09:31 -07:00

model_executor

HiSparse for Sparse Attention (#20343)

2026-03-22 23:09:31 -07:00

model_loader

Upgrade transformers==5.3.0 (#17784)

2026-03-18 13:50:43 -07:00

models

Add SGLang CUDA crash API logging inspired by FlashInfer (#20910)

2026-03-22 16:39:40 +08:00

multimodal

Revert "feat(mm)(grpc): compute M-RoPE positions for preprocessed VL inputs" (#20956)

2026-03-19 18:03:04 -07:00

multiplex

Revert "[Scheduler] Decouple maybe_send_health_check_signal from process_batch_result" (#20259)

2026-03-10 01:58:48 -07:00

observability

Simplify server startup output (#20885)

2026-03-19 13:11:37 -07:00

parser

Add Mistral Small 4 (Pixtral) support (#20708)

2026-03-18 14:15:32 -07:00

ray

[Fix] RayEngine multi-node: co-locate rank0 scheduler with Engine and fix CUDA device setting (#20722)

2026-03-19 10:27:16 -07:00

sampling

…

speculative

[Spec][Ngram] 2/N: Rename branch length to max trie depth (#21181)

2026-03-22 23:35:25 -07:00

tokenizer

…

utils

[Bug Fix] Fix non-streaming request abort failure when --enable-metrics is enabled (#20625)

2026-03-22 19:58:49 -07:00

weight_sync

…

constants.py

…

environ.py

ci: remove IS_BLACKWELL env var; auto-detect Blackwell (#21118)

2026-03-22 23:44:48 -07:00

server_args_config_parser.py

…

server_args.py

[Spec][Ngram] 2/N: Rename branch length to max trie depth (#21181)

2026-03-22 23:35:25 -07:00