Fixes to alternating SWA layers in Gemma2 #31775
HybridCache: Flip order of alternating global-attn/sliding-attn layers
a1a3ccd6
HybridCache: Read sliding_window argument from cache_kwargs
5d9679d3
Gemma2Model: Flip order of alternating global-attn/sliding-attn layers
e45fb6e2
Code formatting
5bde080e
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub