Yuhui Liang ca029e816b Fix missing idle-batch handling in prepare_mlp_sync_batch_raw (#25404)

2026-05-24 23:53:54 -07:00

..

scheduler_components

Fix missing idle-batch handling in prepare_mlp_sync_batch_raw (#25404)

2026-05-24 23:53:54 -07:00

async_dynamic_batch_tokenizer.py

[tokenizer] improve non streaming request processing + some small fixes. (#20310)

2026-04-10 15:46:12 -07:00

cache_controller.py

Remove managers' unused fields (#25433)

2026-05-16 09:17:11 +08:00

communicator.py

refactor: extract FanOutCommunicator and use declarative spec table (#22967)

2026-04-16 15:37:19 -07:00

configure_logging.py

[Utils] Make request dump robust to unpicklable server_args and large meta_info (#24767)

2026-05-09 21:41:41 -07:00

data_parallel_controller.py

Bundle Scheduler rank/size fields into a frozen ParallelState (#25444)

2026-05-16 09:23:51 +08:00

detokenizer_manager.py

perf: migrate Req token-id storage to array.array('q') in Scheduler (#25098)

2026-05-22 10:51:07 -07:00

disagg_service.py

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

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

embed_types.py

[Score API] Hoist query placeholder scan and specialize PositionalEmbeds stacking (#23513)

2026-04-29 13:51:53 -07:00

hisparse_coordinator.py

[Refactor] Rename NSA → DSA: user-facing aliases, file/class/import rename (#25821)

2026-05-20 00:18:04 -07:00

io_struct.py

perf: migrate Req token-id storage to array.array('q') in Scheduler (#25098)

2026-05-22 10:51:07 -07:00

mm_utils.py

[VLM] feat: replace small H2D calls with a single one for qwen-vl (#26167)

2026-05-24 18:27:17 +08:00

multi_tokenizer_mixin.py

Remove managers' unused fields (#25433)

2026-05-16 09:17:11 +08:00

multimodal_processor.py

[Feature] Stronger transformers modeling backend with TP, PP, MoE, VLMs, and torch compile (#19163)

2026-04-02 16:02:33 -07:00

overlap_utils.py

compile _resolve_spec_extras gather kernels (#26129)

2026-05-23 02:34:41 -07:00

prefill_delayer.py

[PrefillDelayer] support NCCL all-gather for cross-DP info sync (#24768)

2026-05-09 21:20:03 -07:00

schedule_batch.py

[VLM] avoid extra cuda-ipc staging for preprocessed input (#26096)

2026-05-24 19:48:06 +08:00

schedule_policy.py

perf: migrate Req token-id storage to array.array('q') in Scheduler (#25098)

2026-05-22 10:51:07 -07:00

scheduler_input_blocker.py

Tiny move files to utils folder (#11166)

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

scheduler_pp_mixin.py

perf: migrate Req token-id storage to array.array('q') in Scheduler (#25098)

2026-05-22 10:51:07 -07:00

scheduler_recv_skipper.py

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

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

scheduler.py

[VLM] try to reuse precomputed padded input ids in scheduler instead of padding (#26097)

2026-05-25 10:27:04 +08:00

template_detection.py

Clean up server startup log noise (#26205)

2026-05-24 14:35:15 -07:00

template_manager.py

Clean up server startup log noise (#26205)

2026-05-24 14:35:15 -07:00

tokenizer_control_mixin.py

Support getting checksums in weight checker (#24537)

2026-05-06 22:59:28 +08:00

tokenizer_manager_score_mixin.py

Replace single-line defensive getattrs with direct access (#25435)

2026-05-16 09:18:43 +08:00

tokenizer_manager.py

[observability] add ServerArgs.stat_loggers for pluggable metrics backend (#24610)

2026-05-24 22:41:28 +08:00

tp_worker.py

[core] step 1: route non-spec seq_lens via FutureMap with per-mode bootstrap fixes (#25944)

2026-05-21 20:15:51 -07:00

utils.py

drop FutureIndices wrapper class (#26085)

2026-05-22 02:44:42 -07:00