Class ov::pass::ConvertPagedAttnInputs#
-
class ConvertPagedAttnInputs : public ov::pass::MatcherPass#
Set precision and shape of KV cache in PagedAttn op based runtime options.
-
struct KVCacheConfig#
-
struct KVCacheConfig#
This page is a nightly version. It may be incomplete or faulty in both content and functionality. Go to the most recent official documentation version, 2024.
Site Navigation
Section Navigation