Update app.py

app.py

@@ -16,6 +16,7 @@ model_id = "meta-llama/Llama-3.2-11B-Vision-Instruct"
 model = MllamaForConditionalGeneration.from_pretrained(model_id, torch_dtype=torch.bfloat16, device_map="auto")
 processor = AutoProcessor.from_pretrained(model_id)
 
+
 @spaces.GPU()
 def predict_clothing(images):
     messages = [{"role": "user", "content":
@@ -34,8 +35,9 @@ def predict_clothing(images):
         with torch.no_grad():
             output = model.generate(**inputs, max_new_tokens=32)
 
-        output_reponse = str(processor.decode(output[0])).split('\n')
-        output_texts.append(output_reponse[-1])
+        # .<|eot_id|>
+        output_reponse = str(processor.decode(output[0])).split('\n')[-1]
+        output_texts.append(output_reponse[:-11])
         
     print(output_texts)
     return output_texts
@@ -44,7 +46,8 @@ def predict_clothing(images):
 @spaces.GPU(duration=180)
 def generate_image(img1, img2, img3, height, width, img_guidance_scale, inference_steps, seed, separate_cfg_infer, offload_model,
                    use_input_image_size_as_output, max_input_image_size, randomize_seed, guidance_scale=3):
-    
+
+    print()
     input_images = [img1, img2, img3]
     # Delete None
     input_images = [img for img in input_images if img is not None]
@@ -61,11 +64,12 @@ def generate_image(img1, img2, img3, height, width, img_guidance_scale, inferenc
                     The {topwear} is in <img><|image_2|></img>. The {bottomwear} is in <img><|image_3|></img>."""
     else:
         input_images = None
+
         
-    
     if randomize_seed:
         seed = random.randint(0, 10000000)
 
+    print(text)
     output = pipe(prompt=text, input_images=input_images, height=height, width=width, guidance_scale=guidance_scale,
                   img_guidance_scale=img_guidance_scale, num_inference_steps=inference_steps, separate_cfg_infer=separate_cfg_infer, 
                   use_kv_cache=True, offload_kv_cache=True, offload_model=offload_model, 
@@ -73,6 +77,7 @@ def generate_image(img1, img2, img3, height, width, img_guidance_scale, inferenc
     img = output[0]
     return img
 
+    
 def get_example():
     case = [
             [   "./imgs/test_cases/icl1.jpg",
@@ -101,6 +106,7 @@ def run_for_examples(img1, img2, img3, height, width, img_guidance_scale, seed,
     return generate_image(text, img1, img2, img3, height, width, img_guidance_scale, inference_steps, seed,
                           separate_cfg_infer, offload_model, use_input_image_size_as_output, max_input_image_size, randomize_seed, guidance_scale)
 
+
 description = """
 This is a Virtual Try-On Platform.
 Usage: