4 people b8a6adadfe [Bug][Spec] fix startup crash and reduce CUDA graph memory usage for speculative adaptive (#35275) ...

Co-authored-by: alphabetc1 <2508695655@qq.com>
Co-authored-by: Shuwen Wang <47200617+alphabetc1@users.noreply.github.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>

2026-08-27 00:46:13 -07:00

..

arg_groups

[Feature] Add DeepEPv2 (ElasticBuffer) MoE A2A backend (#35634)

2026-08-26 19:54:33 -07:00

batch_invariant_ops

[NPU]adaptation to support deterministic inference (#21197)

2026-06-09 09:22:09 +08:00

batch_overlap

[Performance] Reduce idle DP work in breakable prefill CUDA graphs (#33871)

2026-08-27 10:09:54 +08:00

beam_search

[Feature] Beam search support (#31626)

2026-08-26 16:56:15 -07:00

checkpoint_engine

…

compilation

[Memory] Borrow CUDA graph pool storage for EAGLE sampling (#35375)

2026-08-18 16:19:41 -07:00

configs

[Model] Support Ling-3.0-flash (BailingMoeV3) (#33561)

2026-08-26 17:27:23 -07:00

connector

…

constrained

fix(constrained): reject NUL bytes in grammar specs to stop an xgrammar segfault (#34679)

2026-08-19 15:31:14 -07:00

debug_utils

spec: size the speculative buffers from the bags, not the startup record (#35024)

2026-08-17 16:16:56 -07:00

disaggregation

config: the runtime readers take the published bags (#36254)

2026-08-26 05:08:25 -07:00

distributed

[NVIDIA] Tune custom all reduce v2 for sm_107 (#36397)

2026-08-26 15:19:31 -07:00

dllm

config: resolution reads the declarations, not the fields (#36253)

2026-08-26 05:05:28 -07:00

elastic_ep

config: stop handing the record to code that does not read it (#36252)

2026-08-26 05:02:17 -07:00

entrypoints

config: ServerArgs holds the raw input (#36255)

2026-08-26 05:14:05 -07:00

eplb

config: the runtime readers take the published bags (#36254)

2026-08-26 05:08:25 -07:00

function_call

[Model] Support Ling-3.0-flash (BailingMoeV3) (#33561)

2026-08-26 17:27:23 -07:00

hardware_backend

[Model] Support Ling-3.0-flash (BailingMoeV3) (#33561)

2026-08-26 17:27:23 -07:00

kv_canary

config: project the config bags from the resolution result (#35906)

2026-08-23 01:18:24 -07:00

layers

[AMD] Optimize Qwen3.5 MTP unified attention on gfx950 (#36330)

2026-08-27 00:13:49 -07:00

lora

config: resolution reads the declarations, not the fields (#36253)

2026-08-26 05:05:28 -07:00

managers

[Model] Support Ling-3.0-flash (BailingMoeV3) (#33561)

2026-08-26 17:27:23 -07:00

mem_cache

Pin scheduler metadata before asynchronous H2D copies (#35944)

2026-08-26 22:21:47 -07:00

model_executor

[Bug][Spec] fix startup crash and reduce CUDA graph memory usage for speculative adaptive (#35275)

2026-08-27 00:46:13 -07:00

model_loader

[Model] Support Ling-3.0-flash (BailingMoeV3) (#33561)

2026-08-26 17:27:23 -07:00

models

[Feature] Add DeepEPv2 (ElasticBuffer) MoE A2A backend (#35634)

2026-08-26 19:54:33 -07:00

multimodal

[VLM] route every multimodal processor through the worker pool's call site (#35342)

2026-08-27 10:27:45 +08:00

multiplex

config: read resolved config via namespace accessors (#33013)

2026-07-31 15:06:59 -07:00

observability

config: project the config bags from the resolution result (#35906)

2026-08-23 01:18:24 -07:00

parser

[Model] Support Ling-3.0-flash (BailingMoeV3) (#33561)

2026-08-26 17:27:23 -07:00

platforms

[Kernel] Unify BaseFusedOp and MultiPlatformOp dispatch (#33205)

2026-08-06 08:52:09 +08:00

plugins

…

ray

config: ServerArgs holds the raw input (#36255)

2026-08-26 05:14:05 -07:00

rust_extensions

Build Rust extensions on demand in source checkouts (#34994)

2026-08-16 14:58:06 -07:00

sampling

[Feature] Beam search support (#31626)

2026-08-26 16:56:15 -07:00

session

[unified-memory] Stop eviction when shared allocation capacity is sufficient (#33091)

2026-08-26 02:06:32 -07:00

speculative

Pin scheduler metadata before asynchronous H2D copies (#35944)

2026-08-26 22:21:47 -07:00

state_capturer

[Feature] Add DeepEPv2 (ElasticBuffer) MoE A2A backend (#35634)

2026-08-26 19:54:33 -07:00

tokenizer

Add Inkling model support (#31681)

2026-07-19 22:57:37 -07:00

utils

config: the runtime readers take the published bags (#36254)

2026-08-26 05:08:25 -07:00

weight_cache

[Weight Cache] Enhance test and support EPLB (#36198)

2026-08-26 23:11:53 -07:00

weight_sync

Introduce WeightUpdater and WeightExporter components (#31148)

2026-07-14 15:53:33 +08:00

constants.py

…

environ.py

[Feature] Add DeepEPv2 (ElasticBuffer) MoE A2A backend (#35634)

2026-08-26 19:54:33 -07:00

runtime_context.py

Fix _is_compiling dynamo tracing: import torch instead of sys.modules lookup (#36573)

2026-08-26 16:17:02 -07:00

server_args_config_parser.py

fix(server): serialize nested dict config values as JSON (#28094)

2026-06-12 19:27:31 -07:00

server_args.py

[Weight Cache] Enhance test and support EPLB (#36198)

2026-08-26 23:11:53 -07:00