sglang

async_dynamic_batch_tokenizer.py

[Performance] Dynamic Batch Tokenizer (#9382 )

2025-09-14 01:56:04 +08:00

async_mm_data_processor.py

[VLM] Optimize async mm data process mechanism (#12066 )

2025-11-01 01:24:53 +08:00

cache_controller.py

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

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

configure_logging.py

Improve streaming, log_level, memory report, weight loading, and benchmark script (#7632 )

2025-06-29 23:16:19 -07:00

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

feat: TTL-based prefix pinning with refresh-on-hit for HiRadixCache (#18941 )

2026-03-01 23:27:22 -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

Piecewise Cuda Graph set default (#16331 )

2026-03-02 23:18:07 +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

Tiny move files to utils folder (#11166 )

2025-10-03 22:40:06 +08:00

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

Apply default stream to priority 0 in scheduling. (#16438 )

2026-03-03 00:05:27 -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

[Feat]Add scheduler recv skipper weights to environment configuration (#13855 )

2025-11-27 18:16:11 +08:00

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

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

2026-03-03 19:38:15 +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

feat: TTL-based prefix pinning with refresh-on-hit for HiRadixCache (#18941 )

2026-03-01 23:27:22 -08:00

tokenizer_manager_multiitem_mixin.py

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

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

tokenizer_manager.py

Fix routed_dp_rank boundary validation (#19762 )

2026-03-03 11:55:15 -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