Update app.py

app.py

@@ -20,7 +20,7 @@ from prompts import front as front_prompt, back as back_prompt
 # ---------------------------
 # HUGGING FACE MODEL SETUP (CPU)
 # ---------------------------
-path = "OpenGVLab/InternVL2_5-2B"
+path = "OpenGVLab/InternVL2_5-1B"
 cache_folder = "./.cache"
 
 # Load the Vision AI model and tokenizer globally.
@@ -143,7 +143,7 @@ def vision_ai_api(image, doc_type):
     gc.collect()  # Force garbage collection
     torch.cuda.empty_cache()  
     
-    return f'Assistant: {response}'
+    return f'{response}'
 
 # ---------------------------
 # PREDICTION PIPELINE