R0CKSTAR db97f193b7 [diffusion][llm] macOS support (#19549) ...

Signed-off-by: Xiaodong Ye <yeahdongcn@gmail.com>
Co-authored-by: Mick <mickjagger19@icloud.com>

2026-03-10 13:11:07 -07:00

..

batch_invariant_ops

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

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

batch_overlap

Optimization: Reduce the number of D2H operations (#19424)

2026-03-04 16:32:42 +08:00

checkpoint_engine

update pre-commit config (#18860)

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

compilation

Piecewise Cuda Graph set default (#16331)

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

configs

[diffusion][llm] macOS support (#19549)

2026-03-10 13:11:07 -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 grammar sync across TP ranks (#17100)

2026-01-15 18:38:01 -08:00

debug_utils

Enhance error resilience in dump comparator (#19685)

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

disaggregation

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

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

distributed

[ROCm] Use unreg path for aiter custom all-reduce during CUDA graph capture (#20155)

2026-03-09 01:09:04 -07:00

dllm

Decouple scheduler log printing from metrics collection (#20107)

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

elastic_ep

[4/N] (Elastic EP) Back up Expert Weights in DRAM (#17374)

2026-02-27 15:59:13 +08:00

entrypoints

feat(grpc): add SubscribeKvEvents RPC for KV cache event streaming (#20112)

2026-03-08 16:00:29 -07:00

eplb

Refactor: observability code cleanup (#17862)

2026-02-24 18:07:29 -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

grpc

fix(grpc): use context.abort() with proper status codes instead of in-band errors (#19972)

2026-03-05 14:53:18 -08:00

hardware_backend/npu

[NPU] replace swiglu with custom kernel

2026-03-10 21:08:37 +08:00

layers

[diffusion][llm] macOS support (#19549)

2026-03-10 13:11:07 -07:00

lora

[NPU] make torch_native lora backend a little bit faster (#17228)

2026-03-07 20:14:46 +03:00

managers

[diffusion][llm] macOS support (#19549)

2026-03-10 13:11:07 -07:00

mem_cache

[diffusion][llm] macOS support (#19549)

2026-03-10 13:11:07 -07:00

model_executor

[NPU] Bugfix for qwen35 on NPU (#19756)

2026-03-10 20:03:26 +08:00

model_loader

[Quantization] Support config.json quantization_config format, fix exclude_modules matching, and fix KV cache scale loading for Nemotron (#18546)

2026-02-21 16:14:29 +08:00

models

[Qwen-MOE] Fix memory duplication issues in case layers weights are re-assigned during weight loading (#18255)

2026-03-10 17:34:56 +00:00

multimodal

[VLM] Replace decord with torchcodec for video decoding (#20055)

2026-03-09 19:23:49 +08:00

multiplex

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

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

observability

Decouple scheduler log printing from metrics collection (#20107)

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

parser

rename nemotron reasoning parser (#19865)

2026-03-05 11:27:07 -08:00

ray

Add Ray actor support for scheduler process management (DP=1) (#17684)

2026-03-05 13:21:23 -08:00

sampling

Revert "feat: reduce constrained-decoding overhead in TP" (#16845)

2026-01-10 11:39:38 +08:00

speculative

[Bugfix] Skip _mamba_verify_update for idle batch (#20167)

2026-03-10 14:53:01 +08:00

tokenizer

[Piecewise Cuda Graph] rename, refactor and add more logging (#13675)

2025-11-21 13:28:38 +08:00

utils

[diffusion][llm] macOS support (#19549)

2026-03-10 13:11:07 -07:00

weight_sync

[RL] Allow passing tensors of different dtypes for FlattenedTensorBucket (#13413)

2025-11-19 17:24:48 +08:00

constants.py

Support releasing CUDA graph memory when paused (#7873)

2025-10-28 14:40:50 +08:00

environ.py

[diffusion][llm] macOS support (#19549)

2026-03-10 13:11:07 -07:00

server_args_config_parser.py

[ConfigArgumentMerger] Improve ConfigArgumentMerger compatibility with external callers (#17051)

2026-01-16 15:32:42 +08:00

server_args.py

[diffusion][llm] macOS support (#19549)

2026-03-10 13:11:07 -07:00