hlu1 9457c049e1 [Qwen3.5] Enable MTP spec_v2 and add test for nvidia/Qwen3.5-397B-A17B-NVFP4 (#19391)

2026-03-04 14:01:25 -08:00

..

4-gpu-models

[Qwen3.5] Enable MTP spec_v2 and add test for nvidia/Qwen3.5-397B-A17B-NVFP4 (#19391)

2026-03-04 14:01:25 -08:00

8-gpu-models

Fix CompressedTensorsMxInt4MoE abstract method and relax GPQA baseline (#19726)

2026-03-02 19:03:21 -08:00

amd

[AMD] [Qwen 3.5 Day 0] Add Qwen 3.5 nightly accuracy tests (#19479)

2026-03-02 19:42:42 -08:00

ascend

[NPU]grok2 model support (#17119)

2026-03-03 10:24:10 +08:00

attention

fix flaky test for test_kda_kernels (#19864)

2026-03-04 22:47:29 +08:00

backends

…

bench_fn

[Benchmark] Fix generated_shared_prefix attribute naming and remove args dependency (#19363)

2026-02-25 18:45:54 -08:00

constrained_decoding

…

core

[Score API][18132] return token usage in Score API response (#18381)

2026-03-03 13:45:35 -08:00

debug_utils

Enhance error resilience in dump comparator (#19685)

2026-03-02 19:08:35 +08:00

disaggregation

[HiCache] Re-land spec v2 + decode KV cache offloading compatibility (#19615)

2026-03-01 13:58:31 -08:00

distributed

[model gateway][0/N] router EPD support: add encoder grpc server backend support (#16552)

2026-03-03 19:38:15 +08:00

dllm

…

ep

…

eval

Fix nightly VLM accuracy: gemma3n TP fixes + removal, latency thresholds (#19401)

2026-02-27 14:24:02 -08:00

function_call

[Tool Call] Stream DeepSeek-V3.2 function call parameters in JSON format. (#16091)

2026-03-03 01:46:29 -08:00

hicache

…

kernels

…

layers

Fix mamba2 mixer ci test (#19658)

2026-03-02 14:55:15 +08:00

lora

[TestFix] change LoRA tests to use NVIDIA adapter instead of Nutanix (#19642)

2026-03-02 12:55:41 -08:00

metrics

Refactor: observability code cleanup (#17862)

2026-02-24 18:07:29 -08:00

mla

Piecewise Cuda Graph set default (#16331)

2026-03-02 23:18:07 +08:00

model_loading

Fix/nemotron mtp quantaized (#19433)

2026-03-03 01:07:46 -08:00

models

Piecewise Cuda Graph set default (#16331)

2026-03-02 23:18:07 +08:00

moe

…

openai_server

fix: Properly return abort error for streaming requests if the abort is triggered by scheduler (#19357)

2026-03-03 17:18:15 -08:00

ops

[AMD] Support --enable-aiter-allreduce-fusion on AMD GPUs (#13747)

2026-02-24 23:11:55 -08:00

parser

Add KimiK2Detector with tool interruption support (#19696)

2026-03-03 14:04:49 +08:00

perf

[TestFix] change LoRA tests to use NVIDIA adapter instead of Nutanix (#19642)

2026-03-02 12:55:41 -08:00

piecewise_cuda_graph

Piecewise Cuda Graph set default (#16331)

2026-03-02 23:18:07 +08:00

profiling

Migrate 4-GPU/8-GPU workflow jobs to stage-c and add CI registry decorators (#17299)

2026-01-31 22:37:22 -08:00

quant

[CI] Register test_quant_config_parsing.py in CI suite (#19809)

2026-03-03 16:53:31 -08:00

radix_cache

Piecewise Cuda Graph set default (#16331)

2026-03-02 23:18:07 +08:00

rl

…

rotary

…

sampling

…

scheduler

Fix: Reject requests with a duplicate request ID which can cause server crash/hang (#19035)

2026-03-03 00:33:25 -08:00

sessions

[Session] Gate streaming sessions with --enable-streaming-session and spec v2 guard (#19531)

2026-02-27 18:14:55 -08:00

spec

[AMD] Fix EAGLE3 speculative decoding with aiter attention backend (#19362)

2026-03-03 16:12:13 -08:00

stress

…

tokenizer

…

utils

…

vlm

[vlm][minicpm] support input formats of processor output and embedding (#19614)

2026-03-04 12:11:12 -05:00

test_hybrid_dp_ep_tp_mtp.py

…

test_srt_backend.py

…