app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer

# Load the model and tokenizer (fine-tuned or pre-trained)
model_name = "EleutherAI/gpt-neo-1.3B"  # Replace with your fine-tuned model path
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

# Define the function to transform requests
def transform_request(instruction):
    inputs = tokenizer(instruction, return_tensors="pt", truncation=True)
    outputs = model.generate(**inputs, max_length=100)
    code = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return code

# Create the Gradio interface
interface = gr.Interface(
    fn=transform_request,
    inputs="text",
    outputs="text",
    title="Code Transformer",
    description="Enter an instruction to generate Python code.",
)

if __name__ == "__main__":
    interface.launch()