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

# Load pre-trained model and tokenizer
def load_model(model_name):
    tokenizer = AutoTokenizer.from_pretrained(model_name)
    model = AutoModelForCausalLM.from_pretrained(model_name)
    device = "cuda" if torch.cuda.is_available() else "cpu"
    model = model.to(device)
    return tokenizer, model, device

# Function to generate chat responses
def chat_with_niti(message, history):
    tokenizer, model, device = load_model("facebook/mbart-large-50")
    input_ids = tokenizer.encode(message, return_tensors="pt").to(device)
    output = model.generate(
        input_ids,
        max_length=100,
        temperature=0.7,
        num_return_sequences=1,
        pad_token_id=tokenizer.eos_token_id
    )
    response = tokenizer.decode(output[0], skip_special_tokens=True)
    return response

# Create Gradio chat interface
demo = gr.ChatInterface(
    fn=chat_with_niti,
    title="Niti - Your AI Chatbot",
    description="Ask Niti anything in Hindi, Hinglish, or English!"
)

# Launch the interface
demo.launch()