Update app.py

app.py

@@ -156,6 +156,7 @@ def generate_virtual_try_on(person_image, cloth_image, mask_image, pose_image,cl
     cloth_pure = transform(cloth_image).unsqueeze(0).to(device)
     mask_tensor = to_tensor(mask_image)[:1].unsqueeze(0).to(device)  # Keep only one channel
     pose_tensor = transform(pose_image).unsqueeze(0).to(device)
+    
 
     # Prepare text prompts
     prompt = ["A person wearing the cloth"+cloth_des]  # Example prompt
@@ -197,22 +198,22 @@ def generate_virtual_try_on(person_image, cloth_image, mask_image, pose_image,cl
 
     with torch.no_grad():
         images = pipe(
-            prompt_embeds=prompt_embeds,
-            negative_prompt_embeds=negative_prompt_embeds,
-            pooled_prompt_embeds=pooled_prompt_embeds,
-            negative_pooled_prompt_embeds=negative_pooled_prompt_embeds,
+            prompt_embeds=prompt_embedsto.to(device,torch.float16),
+            negative_prompt_embeds=negative_prompt_embeds.to(device,torch.float16),
+            pooled_prompt_embeds=pooled_prompt_embeds.to(device,torch.float16),
+            negative_pooled_prompt_embeds=negative_pooled_prompt_embeds.to(device,torch.float16),
             num_inference_steps=args.num_inference_steps,
             generator=generator,
             strength=1.0,
-            pose_img=pose_tensor,
-            text_embeds_cloth=prompt_embeds_c,
-            cloth=cloth_pure,
-            mask_image=mask_tensor,
+            pose_img=pose_tensor.to(device,torch.float16),
+            text_embeds_cloth=prompt_embeds_c.to(device,torch.float16),
+            cloth=cloth_pure.to(device,torch.float16),
+            mask_image=mask_tensor.to(device,torch.float16),
             image=(person_tensor + 1.0) / 2.0,
             height=args.height,
             width=args.width,
             guidance_scale=guidance_scale,
-            ip_adapter_image=image_embeds,
+            ip_adapter_image=image_embeds.to(device,torch.float16),
         )[0]
 
     # Convert output image to PIL format for display