Pipe in local_global attention

jackzhxng · facebook-github-bot · commit f5a325827c9c · 2025-05-14T14:59:56.000-07:00
Differential Revision: D74762916
diff --git a/examples/models/llama/export_llama_lib.py b/examples/models/llama/export_llama_lib.py
@@ -719,6 +719,7 @@ def _prepare_for_llama_export(args) -> LLMEdgeManager:
             preq_mode=args.preq_mode,
             preq_group_size=args.preq_group_size,
             preq_embedding_quantize=args.preq_embedding_quantize,
+            local_global_attention=args.local_global_attention,
         )
     )
 

Original file line number	Diff line number	Diff line change
`@@ -719,6 +719,7 @@ def _prepare_for_llama_export(args) -> LLMEdgeManager:`
`719`	`719`	`preq_mode=args.preq_mode,`
`720`	`720`	`preq_group_size=args.preq_group_size,`
`721`	`721`	`preq_embedding_quantize=args.preq_embedding_quantize,`
	`722`	`+ local_global_attention=args.local_global_attention,`
`722`	`723`	`)`
`723`	`724`	`)`
`724`	`725`