[ragged-paged-attn] Use hidden states in kv cache and support any num_kv_head #8851
Refactor kv cache to use hidden states
07d1cfe7
Fix undefined variable
3a891781
Fix kv cache shape in tests
ddbe8056
Rename kv_model_dim to kv_hidden_size
add6be02
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub