[MM][CG] Enable encoder Cudagraph for Step3VL (#42224)

Signed-off-by: JisoLya <523420504@qq.com>
Signed-off-by: Soyaazz <523420504@qq.com>
This commit is contained in:
Soyaazz
2026-05-17 20:19:13 -07:00
committed by GitHub
parent 107210442d
commit 990f49bdcb
8 changed files with 384 additions and 22 deletions
+21
View File
@@ -1594,6 +1594,27 @@ class SupportsEncoderCudaGraph(Protocol):
"""
...
def postprocess_encoder_output(
self,
output: torch.Tensor,
indices: list[int],
per_item_out_tokens: list[int],
dest: dict[int, torch.Tensor] | list[torch.Tensor | None],
clone: bool = False,
batch_mm_kwargs: dict[str, Any] | None = None,
) -> None:
"""
Post-process encoder output, directly call scatter_output_slices by default.
By default, delegates directly to scatter_output_slices.
Override this for models that require additional processing on the raw
encoder output prior to scattering, e.g. Step3-VL, which merges features
according to dynamic patch counts before scattering.
"""
from vllm.model_executor.models.utils import scatter_output_slices
scatter_output_slices(output, indices, per_item_out_tokens, dest, clone)
def prepare_encoder_cudagraph_capture_inputs(
self,
token_budget: int,