Update app.py

app.py

@@ -30,11 +30,15 @@ def askme(symptoms, question):
     prompt = tokenizer.apply_chat_template(messages, template=custom_template, tokenize=False, add_generation_prompt=True)
     inputs = tokenizer(prompt, return_tensors="pt").to(device)  # Ensure inputs are on CUDA device
     outputs = model.generate(**inputs, max_new_tokens=200, use_cache=True)
-    response_text = tokenizer.batch_decode(outputs, skip_special_tokens=True)[0].strip()
+    response_text r.batch_decode(outputs, skip_special_tokens=True)[0].strip()
 
     # Extract only the assistant's response
-    #assistant_response = response_text.split("<|im_start|>assistant")[-1].strip()
-    return response_text
+    start_idx = response_text.find("<|im_start|>assistant")
+    end_idx = response_text.find("<|im_end|>", start_idx)
+    assistant_response = response_text[start_idx + len("<|im_start|>assistant"):end_idx]
+    # Return only one answer
+    answers = assistant_response.split(". ")
+    return answers[0] + "."