Update generate.py

generate.py

@@ -77,7 +77,6 @@ class GenerationService:
                 
                 faces = self.face_app.get(face)
                 if faces:
-                    # Shape of normed_embedding is (512,). .unsqueeze(0) makes it (1, 512)
                     faceid_embed = torch.from_numpy(faces[0].normed_embedding).unsqueeze(0)
                     faceid_all_embeds.append(faceid_embed)
             except Exception as e:
@@ -87,27 +86,23 @@ class GenerationService:
             logger.error("No faces were detected in any of the provided images.")
             return None
         
-        # Stack embeds into a single tensor and calculate the average
-        # Shape of stacked tensor: (num_images, 1, 512)
-        # Shape of average_embedding: (1, 512)
         average_embedding = torch.mean(torch.stack(faceid_all_embeds, dim=0), dim=0)
 
         logger.info("Calling the generation pipeline...")
         try:
-            # FIX: The pipeline expects a 3D or 4D tensor.
-            # We add a "sequence length" dimension of 1.
-            # Shape becomes: (1, 1, 512)
             final_embedding = average_embedding.unsqueeze(0)
 
-            image = self.pipe(
+            # FIX: The pipeline returns an object, not a tuple. Access the .images attribute.
+            pipeline_output = self.pipe(
                 prompt=full_prompt,
                 negative_prompt=negative_prompt,
-                ip_adapter_image_embeds=[final_embedding], # Pass the correctly shaped tensor
+                ip_adapter_image_embeds=[final_embedding],
                 num_inference_steps=40,
                 guidance_scale=7.5,
                 width=512,
                 height=768,
-            ).images[0]
+            )
+            image = pipeline_output.images[0]
 
             # --- Save image locally first ---
             temp_dir = "temp_images"