Update app.py

app.py

@@ -8,7 +8,17 @@ tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float16, device_map="auto")  # 降低内存占用
 
 def generate_sql(user_question, create_table_statements, instructions=""):
-    prompt = f"Generate a SQL query to answer this question: `{user_question}`\n{instructions}\n\nDDL statements:\n{create_table_statements}\n<|eot_id|>"
+    prompt = f"""<|begin_of_text|><|start_header_id|>user<|end_header_id|>
+
+Generate a SQL query to answer this question: `{user_question}`
+{instructions}
+
+DDL statements:
+{create_table_statements}<|eot_id|><|start_header_id|>assistant<|end_header_id|>
+
+The following SQL query best answers the question `{user_question}`:
+```sql
+"""
     
     inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
     outputs = model.generate(**inputs, max_length=150)
@@ -17,6 +27,9 @@ def generate_sql(user_question, create_table_statements, instructions=""):
     return sql_query
 
 
+
+
+
 demo = gr.Interface(
     fn=generate_sql,
     inputs=[