Files

T

zhrrr 1edfd09ffd [Model Runner V2] Use actual batch max_seq_len for attn metadata (#43991 )

Signed-off-by: zhuhaoran <zhuhaoran.zhr@alibaba-inc.com>

2026-06-02 06:07:56 +00:00

metrics

[Model Runner V2] Support num NaNs in logits (#30187 )

2025-12-09 10:00:49 -08:00

[Core] Simplify multimodal masking (#34246 )

2026-04-01 01:18:22 -07:00

model_states

[Model Runner V2] Use actual batch max_seq_len for attn metadata (#43991 )

2026-06-02 06:07:56 +00:00

pool

[Frontend] Re-enable running MaxSim on GPU (#38620 )

2026-04-03 00:03:13 +08:00

sample

[Model Runner V2] Fix prompt logprobs calculation Sizes of tensors must match error (#42778 )

2026-05-18 08:27:21 -07:00

spec_decode

[Model Runner V2] Use actual batch max_seq_len for attn metadata (#43991 )

2026-06-02 06:07:56 +00:00

__init__.py

GPU Model Runner V2 (#25266 )

2025-11-21 08:20:55 -08:00

async_utils.py

[BugFix][MRV2] Fix cuda event reuse race (#39115 )

2026-04-07 00:21:09 +00:00

attn_utils.py

[Bugfix] Use storage_block_size in KV cache reshape for compressed specs (DeepSeek V4) (#43988 )

2026-05-29 11:14:25 -07:00

block_table.py

[ModelRunnerV2][Hybrid model] Support kernel block size in hybrid model (#38831 )

2026-05-28 00:55:55 +00:00

buffer_utils.py

[Perf][3/n] Eliminate GPU<->CPU syncs in attention impls (#41434 )

2026-05-07 19:44:24 -07:00

cp_utils.py

[Model Runner V2] Minor simplification for DCP (#34786 )

2026-02-18 11:04:53 -08:00

cudagraph_utils.py

[MRV2] Support breakable CUDA graph (#44050 )

2026-05-30 09:40:52 -07:00

dp_utils.py

[Refactor] Remove unused dead code (#40640 )

2026-04-25 07:28:18 +08:00

eplb_utils.py

[Bugfix] Unwrap VLM wrappers for EPLB on Model Runner V2 (#42706 )

2026-05-16 07:20:33 +08:00

input_batch.py

[Model Runner V2] support qwen35 / mamba hybrid model (#35520 )

2026-05-07 09:31:05 -07:00

kv_connector.py

[Core] Cleanup KVConnector handling with PP + fix MRV2 (#43732 )

2026-05-28 13:12:03 -07:00

lora_utils.py

[ModelRunner V2] Misc minor simplifications and optimizations (#33467 )

2026-02-01 22:17:14 +00:00

model_runner.py

[Model Runner V2] Support zeroing freshly allocated KV blocks for hybrid + fp8 KVCache (#43990 )

2026-06-02 05:56:53 +00:00

pp_utils.py

[Model Runner V2] A bit more PP simplification (#34766 )

2026-02-17 21:39:07 -08:00

README.md

GPU Model Runner V2 (#25266 )

2025-11-21 08:20:55 -08:00

shutdown.py

[MRV2] Add shutdown() method (#41297 )

2026-05-02 21:06:30 -07:00

states.py

[Model Runner V2] Fix seq_lens_cpu_upper_bound (#42202 )

2026-05-11 10:37:50 -07:00

structured_outputs.py

[ModelRunner V2] Misc minor simplifications and optimizations (#33467 )

2026-02-01 22:17:14 +00:00

warmup.py

[MRV2] Ensure warmup covers prefill path (#40746 )

2026-04-24 01:33:26 +00:00

README.md

[Experimental] Model Runner V2

This directory contains the new model runner which is under active development. Ping Woosuk Kwon for any changes.