Commit Graph
8533 Commits
Author SHA1 Message Date
fzyzcjy e737f61b29 Introduce SchedulerLogprobResultProcessor to own logprob state (#25632) 2026-05-18 18:42:59 +08:00
fzyzcjy cf12070a0f Move idle-metrics logging to SchedulerMetricsReporter (#25631) 2026-05-18 18:42:34 +08:00
fzyzcjy fd97fbb096 Move metrics reporting to SchedulerMetricsReporter and retire metrics mixin (#25630) 2026-05-18 18:42:08 +08:00
fzyzcjy 780d969699 Add SchedulerMetricsReporter and route metrics state through it (#25629) 2026-05-18 18:41:35 +08:00
fzyzcjy 8357d07569 Move queue-load reporting to SchedulerLoadInquirer (#25628) 2026-05-18 18:41:12 +08:00
fzyzcjy bde932cbbb Carve out SchedulerLoadInquirer for queue-load state (#25627) 2026-05-18 18:40:46 +08:00
fzyzcjy 1213277879 Move KV-cache event emission to SchedulerKvEventsPublisher (#25626) 2026-05-18 18:40:20 +08:00
fzyzcjy 0f888442c2 Stand up SchedulerKvEventsPublisher; migrate KV-event state to it (#25625) 2026-05-18 18:39:50 +08:00
fzyzcjy f3dce08283 Move invariant checks to SchedulerInvariantChecker and retire runtime_checker mixin (#25624) 2026-05-18 18:39:24 +08:00
fzyzcjy 8a200464fd Introduce SchedulerInvariantChecker to own invariant-check state (#25623) 2026-05-18 18:38:57 +08:00
fzyzcjy b463740953 Move create_scheduler_watchdog from runtime_checker mixin to scheduler.py (#25622) 2026-05-18 18:38:26 +08:00
fzyzcjy ee392a1e14 Move pool-stats sampling to SchedulerPoolStatsObserver (#25621) 2026-05-18 18:37:50 +08:00
fzyzcjy 9fdf73f393 Add SchedulerPoolStatsObserver and route pool-stats state through it (#25619) 2026-05-18 18:36:36 +08:00
fzyzcjy c58b47bc86 Move PoolStats dataclass to scheduler_components.pool_stats_observer (#25618) 2026-05-18 18:35:27 +08:00
fzyzcjy 95d86e5c43 Move on_idle from runtime_checker mixin into Scheduler (#25617) 2026-05-18 18:34:58 +08:00
fzyzcjy 7851ba09f7 Move weight-update RPC handlers to SchedulerWeightUpdaterManager (#25616) 2026-05-18 18:34:15 +08:00
fzyzcjy 56f27635b8 Carve out SchedulerWeightUpdaterManager for weight-update state (#25615) 2026-05-18 18:33:48 +08:00
fzyzcjy a35690f070 Move profiler controls to SchedulerProfilerManager (#25614) 2026-05-18 18:33:17 +08:00
fzyzcjy b0a511560c Stand up SchedulerProfilerManager; migrate profiler state to it (#25613) 2026-05-18 18:32:46 +08:00
fzyzcjy 4d6eec7b32 Move DP-attention adapter methods to SchedulerDPAttnAdapter (#25612) 2026-05-18 18:32:16 +08:00
fzyzcjy 8f37a8a3f3 Introduce SchedulerDPAttnAdapter to own DP-attention state (#25611) 2026-05-18 18:31:45 +08:00
fzyzcjy 0e9eab19a9 Move request-ingress methods to SchedulerRequestReceiver (#25610) 2026-05-18 18:31:17 +08:00
fzyzcjy e6f3dcd790 Add SchedulerRequestReceiver and route request-ingress state through it (#25609) 2026-05-18 18:30:43 +08:00
fzyzcjy 768d347565 Pre-declare mode-conditional Scheduler fields with explicit defaults (#25608) 2026-05-18 18:29:58 +08:00
fzyzcjy 3e3661fd2f Move build_kv_cache to mem_cache.kv_cache_builder (#25607) 2026-05-18 18:29:30 +08:00
fzyzcjy fed1197474 Reshape init_cache_with_memory_pool to match the future build_kv_cache signature (#25606) 2026-05-18 18:28:58 +08:00
fzyzcjy 60337abe24 Hoist hisparse and decode-offload setup out of init_cache_with_memory_pool (#25605) 2026-05-18 18:28:26 +08:00
fzyzcjy 8692bdd3de Move maybe_register_hicache_draft to mem_cache.kv_cache_builder (#25604) 2026-05-18 18:27:58 +08:00
fzyzcjy 189e0a4240 Decouple _maybe_register_hicache_draft from self (#25603) 2026-05-18 18:27:29 +08:00
fzyzcjy 062f6f7ae8 Move get_draft_kv_pool to mem_cache.kv_cache_builder (#25602) 2026-05-18 18:27:01 +08:00
fzyzcjy d1c04deba7 Decouple _get_draft_kv_pool from self before extraction (#25601) 2026-05-18 18:26:20 +08:00
sogalin_codegen abe2ec2aff [AMD] Enable shared-experts fusion with new KIMI-K2.5-MXFP4 model. (#25390) 2026-05-18 01:30:58 -07:00
Xia WeiwenandMa Mingfei 8d5ed330cc [XPU] Enable qwen3.5 on XPU (#21668)
Co-authored-by: Ma Mingfei <mingfei.ma@intel.com>
2026-05-18 14:59:19 +08:00
YAMY 1f9eda4ea1 Use DeepGEMM BF16 for unquantized DeepEP LL MoE (#25540) 2026-05-17 23:33:23 -07:00
xutizhou 6a21dd20b1 Fix EPLB mapping for TopK paths (#25285) 2026-05-17 23:26:04 -07:00
maocheng23 6ccc5b807d Optimize detokenization without HF decode kwargs (#25309) 2026-05-17 20:37:36 -07:00
Zhonghua Deng 784fe7e99b feat(mimo-v2): add EPD disaggregation support (#24931) 2026-05-18 10:33:16 +08:00
Liangsheng Yin 43e133208a Quiet test_bs_1_speed CI log (#25548) 2026-05-17 19:29:44 -07:00
a080358cac [Refactor] Refactor DeepEP dispatcher (#22822)
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Cheng Wan <54331508+ch-wan@users.noreply.github.com>
2026-05-18 04:36:42 +03:00
chenxu214 2a357071ec [NPU]fix:NPUMLATokenToKVPool object has no attribute "kv_buffer" (#25249) 2026-05-18 09:09:28 +08:00
Liangsheng Yin 58ece60703 refactor: remove ModelWorkerBatch indirection (#25516) 2026-05-17 18:05:39 -07:00
jy-song-hub b3803164cb [diffusion] fix: fix unipc device placement + flowunipc sigma_min crash (#23294) 2026-05-18 09:01:07 +08:00
Shu WangandKhoa Pham c67b287056 Enable trtllm_mha as gemma4 default attn backend. (#25006)
Co-authored-by: Khoa Pham <khoa.pham@radixark.ai>
2026-05-17 14:58:12 -07:00
Cheng WanandClaude Opus 4.7 7158a255eb [MoE Refactor] Migrate flashinfer_cutedsl + DeepEP to MoeRunner (#25525)
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-17 14:48:17 -07:00
Mick 89e501c5a8 [diffusion] CI: tighten selected perf baselines (#25510) 2026-05-17 23:35:59 +08:00
Qingfu WenandR0CKSTAR 3bf7e346fc [MUSA][Diffusion] Improve wan model inference speed using torch.compile (#25256)
Co-authored-by: R0CKSTAR <yeahdongcn@gmail.com>
2026-05-17 22:10:24 +08:00
Mick eccfd6dea7 [diffusion] feat: configure encoder as layerwise-offload by default (#25517) 2026-05-17 20:47:48 +08:00
Cheng WanandClaude Opus 4.7 be3c425788 [MoE] Unify DeepEPMoE+MoriEPMoE through AITER MoeRunner pre/post-permute (#23760)
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-17 02:25:52 -07:00
Cheng Wan 568ba7216a Fix logging for inplace setting in the flashInfer-trtllm backend (#25522) 2026-05-17 02:20:14 -07:00
Yongji WuandZhangheng e547f3f804 fix(unified radix cache w/ hicache): backup ancestor nodes before leaf in write_back eviction (#24585)
Co-authored-by: Zhangheng <hzh0425@apache.org>
2026-05-17 15:58:46 +08:00