Nick Hill e6cb56337b [Core] MRV2 support for fast-prefill (#56145) ...

Signed-off-by: Nick Hill <nickhill123@gmail.com>

2026-09-10 10:36:04 -07:00

..

cpu

[CPU] Ship triton-cpu wheel and fix several hardcoded pin_memory=True (#52092)

2026-08-13 14:40:53 +08:00

gpu

[Core] MRV2 support for fast-prefill (#56145)

2026-09-10 10:36:04 -07:00

sentinel

[Feature] Add fault tolerance framework (simplified) for DP+EP external LB deployments (#44428)

2026-07-25 11:49:09 -04:00

__init__.py

…

block_table.py

[2/N][warmup][DSv4] Migrate sequence and DCP kernels (#53564)

2026-09-07 09:35:54 -04:00

cp_utils.py

[Refactor] Remove utils dead code (#53941)

2026-09-08 11:22:09 -04:00

cpu_model_runner.py

[Bugfix][MRV2] Isolate sleep-mode KV allocations (#53508)

2026-08-27 15:08:28 -07:00

cpu_worker.py

[CI/Build][CPU] Fix flaky rust downloads, broken prune flag, and triton-cpu cache coupling (#56247)

2026-09-10 22:27:26 +08:00

dp_utils.py

[Perf] Ensure async h2d copies are pinned in more places (#55202)

2026-09-04 17:17:01 -07:00

ec_connector_model_runner_mixin.py

[EPD] Add ECMooncakeConnector for encoder cache over Mooncake TransferEngine (#41567)

2026-09-07 06:21:43 +00:00

encoder_cudagraph_defs.py

[MM][CG] Enable encoder CUDA Graph for MiniCPM-V (#42785)

2026-09-10 06:28:48 +00:00

encoder_cudagraph.py

[MM][CG] Enable encoder CUDA Graph for MiniCPM-V (#42785)

2026-09-10 06:28:48 +00:00

gpu_input_batch.py

[Core][Model] Unify XD-RoPE into M-RoPE and derive the channel count (#56078)

2026-09-09 09:23:30 -07:00

gpu_model_runner.py

[Core][Model] Unify XD-RoPE into M-RoPE and derive the channel count (#56078)

2026-09-09 09:23:30 -07:00

gpu_ubatch_wrapper.py

[1/2][Model Runner V2] DBO support, eager mode (#50945)

2026-09-04 14:17:39 -07:00

gpu_worker.py

[EPD] Add ECMooncakeConnector for encoder cache over Mooncake TransferEngine (#41567)

2026-09-07 06:21:43 +00:00

kv_connector_model_runner_mixin.py

[Core][KV Connector] Start async KV loads after the forward launch when no sync loads are scheduled (#53333)

2026-08-28 21:58:00 +00:00

lora_model_runner_mixin.py

[Bugfix] Fix level-2 sleep/wake/reload with enable_lora=True (#39935)

2026-08-06 09:13:10 +08:00

mamba_utils.py

[K3 Bug] Fix Kimi-K3 RecoverSSM startup failure 'MambaAttentionBackendEnum.GDN_ATTN declares 4 states, but provides 2 state copy funcs' (#54634)

2026-08-31 23:22:49 +00:00

mm_encoder_model_runner.py

[Bugfix][V2] Warm up kernels before capturing CUDA graphs (#55341)

2026-09-04 22:44:18 +00:00

startup_plan.py

[Core] Persist and reuse the memory-profiling result across boots (opt-in) (#47388)

2026-07-08 00:53:02 +00:00

tpu_input_batch.py

fix: improve token_ids_cpu swap to copy only valid indices (#36255)

2026-08-28 17:22:39 -07:00

ubatch_utils.py

[1/2][Model Runner V2] DBO support, eager mode (#50945)

2026-09-04 14:17:39 -07:00

ubatching.py

[Misc] Replace Optional[X] with X | None syntax (#33332)

2026-01-30 01:56:59 -08:00

utils.py

[Model] add GLM-5.3-Flash support (#53906)

2026-09-03 09:40:35 -07:00

worker_base.py

[Bugfix][DP] Synchronize the device on pause completion (#52914)

2026-08-26 09:50:33 -07:00

workspace.py

[Bugfix] Make DSV4 sparse MLA work end-to-end for plain decode, MTP, and DSpark (#51538)

2026-08-15 07:01:32 -07:00

xpu_model_runner.py

[XPU] Alias is_current_stream_capturing to XPU in cuda wrapper (#50526)

2026-08-05 15:10:17 +08:00

xpu_worker.py

[XPU] Fix device assignment for DP external LB (#53037)

2026-09-04 15:51:25 +08:00