import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained("jsbeaudry/makandal-pre-trained")
model = AutoModelForCausalLM.from_pretrained("jsbeaudry/makandal-pre-trained")

# Set device
device = "cuda" if torch.cuda.is_available() else "cpu"
model.to(device)

# Generation function
def generate_text(prompt):
    inputs = tokenizer(prompt, return_tensors="pt", padding=True).to(device)
    output = model.generate(
        **inputs,
        max_new_tokens=30,
        do_sample=True,
        repetition_penalty=1.2,
        no_repeat_ngram_size=3,
        temperature=0.9,
        top_k=40,
        top_p=0.85,
        pad_token_id=tokenizer.pad_token_id,
        eos_token_id=tokenizer.eos_token_id
    )
    return tokenizer.decode(output[0], skip_special_tokens=True)

# Gradio interface
iface = gr.Interface(
    fn=generate_text,
    inputs=gr.Textbox(lines=2, placeholder="Ekri yon sijè oswa yon fraz..."),
    outputs="text",
    title="Makandal Text Generator",
    description="Ekri yon fraz oswa mo kle pou jenere tèks ak modèl Makandal la. Modèl sa fèt espesyalman pou kontèks Ayiti."
)

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