Create app.py

app.py

@@ -0,0 +1,152 @@
+import gradio as gr
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import torch
+
+print("🚀 Loading model...")
+
+# Load your merged model
+model_name = "Abhisek987/llama-3.2-sql-merged"
+
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    device_map="auto",
+    torch_dtype=torch.float16,
+    low_cpu_mem_usage=True
+)
+
+print("✅ Model loaded successfully!")
+
+def generate_sql(database, question):
+    """Generate SQL query from natural language question"""
+    
+    prompt = f"""### Instruction:
+You are a SQL expert. Generate a SQL query to answer the given question for the specified database.
+
+### Input:
+Database: {database}
+Question: {question}
+
+### Response:
+"""
+    
+    # Tokenize and generate
+    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+    
+    with torch.no_grad():
+        outputs = model.generate(
+            **inputs,
+            max_new_tokens=256,
+            temperature=0.1,
+            do_sample=True,
+            pad_token_id=tokenizer.eos_token_id,
+            eos_token_id=tokenizer.eos_token_id
+        )
+    
+    # Decode and extract SQL
+    result = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    sql_query = result.split("### Response:")[-1].strip()
+    
+    return sql_query
+
+# Create Gradio interface
+demo = gr.Interface(
+    fn=generate_sql,
+    inputs=[
+        gr.Textbox(
+            label="Database Name",
+            placeholder="e.g., employees, sales, customers",
+            value="employees"
+        ),
+        gr.Textbox(
+            label="Question",
+            placeholder="e.g., Show all employees with salary above 60000",
+            lines=3
+        )
+    ],
+    outputs=gr.Textbox(
+        label="Generated SQL Query",
+        lines=5
+    ),
+    title="🤖 Text-to-SQL Generator",
+    description="""
+    Fine-tuned Llama 3.2 3B model for SQL query generation using LoRA.
+    
+    Enter a database name and your question in natural language, and the model will generate the corresponding SQL query.
+    """,
+    examples=[
+        ["employees", "Show all employees with salary above 60000"],
+        ["sales", "Show me the top 5 products by total sales"],
+        ["customers", "How many customers are from each country?"],
+        ["orders", "Find all orders placed in the last 30 days"],
+        ["concert_singer", "What are the names of all singers ordered by net worth?"]
+    ],
+    theme="soft",
+    article="""
+    ### About This Model
+    
+    - **Base Model:** Llama 3.2 3B
+    - **Fine-tuning Method:** LoRA (Parameter-Efficient Fine-Tuning)
+    - **Dataset:** Spider (7,000 training examples)
+    - **Training Loss:** 0.37 (85% reduction)
+    
+    [View Model on HuggingFace →](https://huggingface.co/Abhisek987/llama-3.2-sql-lora)
+    """
+)
+
+if __name__ == "__main__":
+    demo.launch()
+```
+
+5. Click **"Commit new file to main"**
+
+---
+
+### **Step 3: Create `requirements.txt` File**
+
+1. Click **"Add file"** → **"Create a new file"**
+2. Filename: `requirements.txt`
+3. Paste this:
+```
+transformers>=4.44.0
+torch>=2.0.0
+gradio>=4.0.0
+accelerate>=0.30.0
+```
+
+4. Click **"Commit new file to main"**
+
+---
+
+### **Step 4: Wait for Build**
+
+1. Go to **"App"** tab
+2. You'll see: "Building..." (takes 5-10 minutes)
+3. Once done, your app will be live!
+
+---
+
+### **Step 5: Your Public URL**
+
+Once built, your app will be available at:
+```
+https://huggingface.co/spaces/Abhisek987/sql-generator
+```
+
+---
+
+## **What You'll Get:**
+
+✅ **Free forever** (no charges!)
+✅ **Public demo URL** for resume
+✅ **Works on any device** (runs on HF servers)
+✅ **Professional interface**
+✅ **Example queries** for visitors to try
+✅ **No maintenance needed**
+
+---
+
+## **For Your Resume:**
+```
+🔗 Live Demo: https://huggingface.co/spaces/Abhisek987/sql-generator
+🔗 Model: https://huggingface.co/Abhisek987/llama-3.2-sql-lora