Update handler.py

handler.py

@@ -121,7 +121,7 @@ class EndpointHandler:
                         guidance_scale=guidance_scale,
                         num_inference_steps=num_inference_steps,
                         output_type="pt"
-                    ).images
+                    ).frames[0]
                 else:
                     # Generate video from text only
                     output = self.text_to_video(
@@ -130,7 +130,7 @@ class EndpointHandler:
                         guidance_scale=guidance_scale,
                         num_inference_steps=num_inference_steps,
                         output_type="pt"
-                    ).images
+                    ).frames[0]
 
                 # Convert frames to video file
                 video_content = self._create_video_file(output, fps=fps)