Update handler.py

handler.py

@@ -174,7 +174,7 @@ class EndpointHandler:
             # same type of inference operations multiple times.
             # For example, calling the compiled pipeline on a different image size
             # triggers compilation again which can be expensive.
-            self.image_to_video.unet = torch.compile(self.image_to_video.unet, mode="reduce-overhead", fullgraph=True)
+            self.image_to_video = torch.compile(self.image_to_video, mode="reduce-overhead", fullgraph=True)
 
         else:
             print("EndpointHandler.__init__(): initializing LTXPipeline..")
@@ -193,7 +193,7 @@ class EndpointHandler:
             # same type of inference operations multiple times.
             # For example, calling the compiled pipeline on a different image size
             # triggers compilation again which can be expensive.
-            self.text_to_video.unet = torch.compile(self.text_to_video.unet, mode="reduce-overhead", fullgraph=True)
+            self.text_to_video = torch.compile(self.text_to_video, mode="reduce-overhead", fullgraph=True)
 
 
         # Initialize LoRA tracking