import gradio as gr
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

# Load the fine-tuned model and tokenizer
model = AutoModelForSeq2SeqLM.from_pretrained("Codellama-7b-Instruct")
tokenizer = AutoTokenizer.from_pretrained("Codellama-7b-Instruct")

# Define a function to generate a response from the model
def generate_response(input_text):
    inputs = tokenizer(input_text, return_tensors="pt")
    outputs = model.generate(**inputs)
    response = tokenizer.decode(outputs[0])
    return response

# Create a Gradio interface
interface = gr.Interface(generate_response, input_type="text", output_type="text",
                         title="Codellama-7b-Instruct Chatbot",
                         description="A chatbot powered by the Codellama-7b-Instruct model.",
                         article="This chatbot is fine-tuned on a dataset of instructional text and can be used to generate responses to natural language prompts.",
                         theme="default",
                         share=True,
                         enable_chat=True)

# Launch the interface on a local server
interface.launch()