Liangsheng Yin 5acb45cf32 [Session] Extract SessionController and clean up session logic in Scheduler (#19547)

2026-02-28 19:47:44 -08:00

..

async_dynamic_batch_tokenizer.py

…

async_mm_data_processor.py

…

cache_controller.py

[HiCache] Support heterogeneous tp for hicache storage (#18541)

2026-02-25 00:13:57 -08:00

configure_logging.py

…

data_parallel_controller.py

[PD-Disagg] Fully support external DP dispatch w/ PD-disaggregation mode. (#19268)

2026-02-24 19:58:01 -08:00

detokenizer_manager.py

[PD-Disagg] Fully support external DP dispatch w/ PD-disaggregation mode. (#19268)

2026-02-24 19:58:01 -08:00

disagg_service.py

[PD] Cleanup BootstrapServer init and ready check (#19551)

2026-02-28 16:41:42 +08:00

io_struct.py

[Fix] Add --disable-draft-model-update to control draft model updates(especially in RL) (#15726)

2026-02-28 12:09:55 -08:00

mm_utils.py

model: support Qwen3.5 (#18489)

2026-02-10 00:27:59 +08:00

multi_tokenizer_mixin.py

[PD-Disagg] Fully support external DP dispatch w/ PD-disaggregation mode. (#19268)

2026-02-24 19:58:01 -08:00

multimodal_processor.py

Skip mm feature pool init to avoid EPD OOM (#16388)

2026-01-23 10:53:45 +08:00

overlap_utils.py

[spec v2]Fix torch gc of future indices (#18958)

2026-02-19 11:38:25 -08:00

prefill_delayer.py

Added the prefill delayer policy: The prefill deplay range is expanded. (#17456)

2026-02-28 08:56:49 +08:00

schedule_batch.py

Revert "[SGL] sync patch: Remove sync points, prefill cudagraph for DP, disable cache reset in mem check (#19190)" (#19581)

2026-02-28 19:46:47 -08:00

schedule_policy.py

[NPU] bugs fix for Deepseek models (#19544)

2026-02-28 17:26:15 +08:00

scheduler_dp_attn_mixin.py

Refactor: observability code cleanup (#17862)

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

scheduler_input_blocker.py

…

scheduler_output_processor_mixin.py

[Session] Add streaming mode with SessionAwareCache fast path (#19171)

2026-02-27 16:31:08 -08:00

scheduler_pp_mixin.py

[PD] Support PD with context parallel after refactor (#19504)

2026-02-28 13:11:15 +08:00

scheduler_profiler_mixin.py

[NPU] torch_npu profiler tensorboard path type fix (#17545)

2026-01-24 22:55:49 +08:00

scheduler_recv_skipper.py

…

scheduler_runtime_checker_mixin.py

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

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

scheduler_update_weights_mixin.py

[Fix] Add --disable-draft-model-update to control draft model updates(especially in RL) (#15726)

2026-02-28 12:09:55 -08:00

scheduler.py

[Session] Extract SessionController and clean up session logic in Scheduler (#19547)

2026-02-28 19:47:44 -08:00

session_controller.py

[Session] Extract SessionController and clean up session logic in Scheduler (#19547)

2026-02-28 19:47:44 -08:00

template_manager.py

fix: fix the bug blocking completion template application (#17010)

2026-02-24 18:50:08 +08:00

tokenizer_communicator_mixin.py

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

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

tokenizer_manager_multiitem_mixin.py

feat: support qwen3(-VL) rerank scoring&chat template (#16403)

2026-01-15 00:45:46 +08:00

tokenizer_manager.py

[PD-Disagg] Fully support external DP dispatch w/ PD-disaggregation mode. (#19268)

2026-02-24 19:58:01 -08:00

tp_worker.py

update pre-commit config (#18860)

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

utils.py

Introduce global alloc_len_per_decode & clean check decode memory (#15115)

2026-01-26 10:26:20 -08:00