app.py

import gradio as gr
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

model_name = "t5-base"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)

def generate_text(input_text):
    # Preprocess input text
    input_text = input_text.strip()
    
    # Prepare input for the model
    input_ids = tokenizer.encode("humanize: " + input_text, return_tensors="pt", max_length=512, truncation=True)
    
    # Generate text with improved parameters
    outputs = model.generate(
        input_ids,
        max_length=300,
        min_length=30,
        num_return_sequences=1,
        no_repeat_ngram_size=3,
        top_k=50,
        top_p=0.95,
        temperature=0.8,
        do_sample=True,
        early_stopping=True
    )
    
    # Decode and clean up the generated text
    generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return generated_text.strip()

iface = gr.Interface(
    fn=generate_text,
    inputs=gr.Textbox(lines=5, label="Input Text"),
    outputs=gr.Textbox(label="Generated Text"),
    title="Text Generator",
    description="Enter text to generate a summary or continuation."
)

iface.launch()