Update app.py

app.py

@@ -1,4 +1,5 @@
 # LoRA Inference Gradio Space Demo
+
 import spaces
 import gradio as gr
 from peft import PeftModel
@@ -22,16 +23,20 @@ tokenizer = AutoTokenizer.from_pretrained("unsloth/qwen2.5-3b-instruct-unsloth-b
 
 @spaces.GPU
 def generate_response(prompt):
-    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
-    outputs = model.generate(**inputs, max_new_tokens=50)
+    reasoning_prompt = (
+        "Answer the following question and explain your reasoning step by step.\n"
+        f"Question: {prompt}\nReasoning:"
+    )
+    inputs = tokenizer(reasoning_prompt, return_tensors="pt").to(model.device)
+    outputs = model.generate(**inputs, max_new_tokens=150)
     return tokenizer.decode(outputs[0], skip_special_tokens=True)
 
 zk_qwen = gr.Interface(
     fn=generate_response,
     inputs=gr.Textbox(lines=2, placeholder="Enter your prompt here..."),
     outputs=gr.Textbox(label="Response"),
-    title="LoRA Model Inference",
-    description="Demo your LoRA model with Hugging Face Gradio."
+    title="LoRA Model Reasoning Inference",
+    description="Demo your LoRA model with step-by-step reasoning in Hugging Face Gradio."
 )
 
 zk_qwen.launch()