app.py

import gradio as gr
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM  # or AutoModelForCausalLM, etc.

# Replace this with the actual path to your model on HF hub
MODEL_NAME = "AbdulrahmanCS/my_finetuned_model"

# Load the tokenizer and model
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
model = AutoModelForSeq2SeqLM.from_pretrained(MODEL_NAME)

# You can wrap it in a pipeline, or do manual model calls.
# For demonstration, let’s create a small function that does text generation:

def generate_interview_questions(prompt):
    inputs = tokenizer(prompt, return_tensors="pt")
    outputs = model.generate(
        **inputs,
        max_length=128,
        num_beams=4,
        num_return_sequences=1,
    )
    return tokenizer.decode(outputs[0], skip_special_tokens=True)

# Now define the Gradio interface
with gr.Blocks() as demo:
    gr.Markdown("# My Fine-Tuned Interview Question Generation Model")

    with gr.Row():
        input_text = gr.Textbox(label="Enter prompt or context", lines=3)
        output_text = gr.Textbox(label="Generated question(s)")

    generate_button = gr.Button("Generate")
    generate_button.click(fn=generate_interview_questions, inputs=input_text, outputs=output_text)

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