app.py

import gradio as gr
from transformers import AutoTokenizer, PegasusForConditionalGeneration

# Load the Humaneyes model from Hugging Face
tokenizer = AutoTokenizer.from_pretrained("Eemansleepdeprived/Humaneyes")
model = PegasusForConditionalGeneration.from_pretrained("Eemansleepdeprived/Humaneyes")

# Ensure the model has a pad_token_id (use eos_token_id if missing)
if model.config.pad_token_id is None:
    model.config.pad_token_id = tokenizer.eos_token_id

def humanize_text(ai_text):
    if not ai_text.strip():
        return "❌ Please enter some text to process."
    # Tokenize the input text
    inputs = tokenizer(ai_text, return_tensors="pt")
    # Generate human-like text with controlled parameters
    outputs = model.generate(
        inputs["input_ids"],
        max_length=256,
        num_beams=5,
        early_stopping=True
    )
    # Decode and return the generated text
    return tokenizer.decode(outputs[0], skip_special_tokens=True)

# Build a modern Gradio UI
with gr.Blocks(theme=gr.themes.Soft(), css=".container {max-width: 700px; margin: auto;}") as demo:
    gr.Markdown("# ✨ AI to Human Text Converter ✨")
    gr.Markdown("Convert AI-generated text into natural, human-like text!")
    
    with gr.Row():
        ai_input = gr.Textbox(
            label="Enter AI Text", 
            placeholder="Type or paste AI-generated text here...", 
            lines=7
        )
    
    btn = gr.Button("🚀 Humanize Text", variant="primary")
    
    with gr.Row():
        human_output = gr.Textbox(
            label="Humanized Output", 
            interactive=False, 
            lines=7
        )
    
    btn.click(humanize_text, inputs=ai_input, outputs=human_output)

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