enableBlockReuse is available for multimodal ?

[YSF-A]

Is there a way to use enableBlockReuse/prefix_cache for multimodal?
I try to set enableBlockReuse=True in MultiModalRunner.py when create ModelRunnerCpp and it raise error.
I do not know if the enableBlockReuse is available for multimodal?

[karljang]

@YSF-A , apologies for the very delayed response.
PR #5444 introduced support for KV cache reuse in multimodal models. So the feature might be available, though it ultimately depends on your specific model architecture and configuration.

[github-actions]

Issue has not received an update in over 14 days. Adding stale label.

[github-actions]

This issue was closed because it has been 14 days without activity since it has been marked as stale.