Update app.py

app.py

@@ -39,7 +39,7 @@ Response:
     if model.config.decoder_start_token_id is None:
         model.config.decoder_start_token_id = tokenizer.pad_token_id
 
-    # Generate SQL output with no_grad to optimize CPU usage.
+    # Generate SQL output using no_grad for optimized CPU usage.
     with torch.no_grad():
         generated_ids = model.generate(
             input_ids=inputs["input_ids"],
@@ -56,7 +56,29 @@ Response:
     generated_sql = generated_sql.split(";")[0].strip() + ";"  # Keep only the first valid SQL query
     return generated_sql
 
-# Create Gradio interface with two input boxes: one for context and one for query.
+# Guide text with detailed instructions and an example.
+guide_text = """
+# Text-to-SQL Inference App Guide
+**Overview:**
+This app uses a fine-tuned FLAN-T5 model to generate SQL queries based on your inputs.
+
+**How to Use:**
+- **Context:** Enter your database schema (table definitions, DDL statements, sample data).
+- **Query:** Enter a natural language query describing the desired SQL operation.
+- Click **Generate SQL** to see the model-generated SQL query.
+
+**Example:**
+- **Context:**
+CREATE TABLE students (id INT PRIMARY KEY, name VARCHAR(100), age INT, grade CHAR(1)); INSERT INTO students (id, name, age, grade) VALUES (1, 'Alice', 14, 'A'), (2, 'Bob', 15, 'B');
+
+- **Query:**
+Retrieve the names of students who are 15 years old.
+
+The generated SQL might look like:
+SELECT name FROM students WHERE age = 15;
+"""
+
+# Create Gradio interface.
 iface = gr.Interface(
     fn=generate_sql,
     inputs=[
@@ -65,9 +87,10 @@ iface = gr.Interface(
     ],
     outputs="text",
     title="Text-to-SQL Generator",
-    description="Enter your own context (e.g., database schema and sample data) and a natural language query. The model will generate the corresponding SQL statement.",
+    description=guide_text,
     theme="compact",
     allow_flagging="never"
 )
 
 iface.launch()
+