add to mps device in _build_GOT_vision

got_vision_b.py

@@ -448,6 +448,8 @@ def _build_GOT_vision(
     image_size = 1024
     vit_patch_size = 16
     image_embedding_size = image_size // vit_patch_size
+    device = torch.device('mps')
+
     image_encoder=ImageEncoderViT(
             depth=encoder_depth,
             embed_dim=encoder_embed_dim,
@@ -461,8 +463,8 @@ def _build_GOT_vision(
             global_attn_indexes=encoder_global_attn_indexes,
             window_size=14,
             out_chans=prompt_embed_dim,
-        )
-    
+        ).to(device)
+
 
     return image_encoder