solve llama3.2 import issue

Author: Gasoonjia

torchchat/model.py

@@ -31,7 +31,7 @@
 )
 from torch.nn import functional as F
 
-from torchtune.models.flamingo import flamingo_decoder, flamingo_vision_encoder
+from torchtune.models.llama3_2_vision import llama3_2_vision_decoder, llama3_2_vision_encoder
 from torchtune.models.llama3_1._component_builders import llama3_1 as llama3_1_builder
 from torchtune.modules.model_fusion import DeepFusionModel
 from torchtune.models.clip import clip_vision_encoder
@@ -213,7 +213,7 @@ def _llama3_1(cls):
     def _flamingo(cls):
         return cls(
             model_type=ModelType.Flamingo,
-            modules={"encoder": flamingo_vision_encoder, "decoder": flamingo_decoder},
+            modules={"encoder": llama3_2_vision_encoder, "decoder": llama3_2_vision_decoder},
             fusion_class=DeepFusionModel,
         )