app.py

import streamlit as st
from transformers import AutoModelForCausalLM, AutoTokenizer

# Set page configuration
st.set_page_config(page_title="SQLCoder", layout="wide")

# Load the model and tokenizer
@st.cache_resource
def load_model():
    model_name = "defog/sqlcoder"
    tokenizer = AutoTokenizer.from_pretrained(model_name)
    model = AutoModelForCausalLM.from_pretrained(
        model_name,
        device_map="auto",
        load_in_8bit=True,  # Use int8 quantization
        low_cpu_mem_usage=True,
    )
    return tokenizer, model

# @st.cache_resource
# def load_model():
#     model_name = "defog/sqlcoder-7b-2"
#     tokenizer = AutoTokenizer.from_pretrained(model_name)
#     model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto")
#     return tokenizer, model

tokenizer, model = load_model()

# Define the function to generate code
def generate_code(prompt, max_length=150):
    inputs = tokenizer(prompt, return_tensors="pt")
    outputs = model.generate(**inputs, max_length=max_length, num_return_sequences=1)
    return tokenizer.decode(outputs[0], skip_special_tokens=True)

# Streamlit app layout
st.title("SQLCoder: AI-Powered SQL Code Generator")
st.write("Generate SQL queries and code snippets using the SQLCoder model.")

# Input for user prompt
prompt = st.text_area("Enter your query or code prompt:", height=150)
max_length = st.slider("Maximum Output Length", min_value=50, max_value=300, value=150, step=10)

# Generate button
if st.button("Generate Code"):
    if prompt.strip():
        with st.spinner("Generating code..."):
            try:
                generated_code = generate_code(prompt, max_length=max_length)
                st.text_area("Generated Code:", value=generated_code, height=200)
            except Exception as e:
                st.error(f"Error generating code: {e}")
    else:
        st.warning("Please enter a prompt before generating code.")