app.py

import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline

# Load model and tokenizer only once
model_name = "EmTpro01/codellama-Code-Generator"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", load_in_8bit=True)

# Create pipeline once
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

def generate_response(prompt):
    # Use the pre-loaded pipeline
    response = pipe(prompt, max_length=1024, temperature=0.7, top_p=0.95, repetition_penalty=1.15)
    return response[0]['generated_text']

iface = gr.Interface(
    fn=generate_response,
    inputs="text",
    outputs="text",
    title="Code Generation Model"
)

iface.launch()