app.py

import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM

model_id = "Saibalaji25/autotrain-0u37b-accmn"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

def generate_code(prompt, max_tokens=128, temperature=0.7, top_p=0.95):
    inputs = tokenizer(prompt, return_tensors="pt")
    outputs = model.generate(
        **inputs,
        max_new_tokens=max_tokens,
        do_sample=True,
        temperature=temperature,
        top_p=top_p,
        pad_token_id=tokenizer.eos_token_id
    )
    return tokenizer.decode(outputs[0], skip_special_tokens=True)

iface = gr.Interface(
    fn=generate_code,
    inputs=[
        gr.Textbox(label="Enter a code prompt", lines=4, placeholder="e.g. def hello_world():"),
        gr.Slider(32, 512, value=128, label="Max Tokens"),
        gr.Slider(0.1, 1.5, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(0.5, 1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)")
    ],
    outputs=gr.Code(label="Generated Code"),
    title="🔧 CodeGen-350M Demo",
    description="A fine-tuned code generation model based on Salesforce/codegen-350M-mono. Enter a function or code comment and generate completions!"
)

iface.launch()