[Perf][DSv4.1] Run decoder replay layers in CUDA graphs and trim in PIECEWISE graphs (#59532)

Signed-off-by: Juntian Liu <Juntianl777@gmail.com>
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
This commit is contained in:
Juntian Liu
2026-10-05 22:11:45 -07:00
committed by GitHub
co-authored by Claude Opus 5.5 mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
parent 21d93d0d8c
commit eb2901b735
10 changed files with 445 additions and 71 deletions
+3 -1
View File
@@ -259,7 +259,9 @@ def _make_capture_runner(captured: bool) -> GPUModelRunner:
"""Minimal V2 runner for capture_model: fakes everything except the
cudagraph_manager's needs_capture decision."""
runner = GPUModelRunner.__new__(GPUModelRunner)
runner.model_state = SimpleNamespace(supports_mm_inputs=False)
runner.model_state = SimpleNamespace(
supports_mm_inputs=False, capture_inner_cudagraphs=lambda *args: None
)
runner.cudagraph_manager = SimpleNamespace(
needs_capture=lambda: captured,
capture=lambda *args, **kwargs: None,