app.py

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

# Load Llama-2 model
model_name = "meta-llama/Llama-2-7b-chat-hf"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float16, device_map="auto")

# Define personalities
personalities = {
    "Albert Einstein": "You are Albert Einstein, the famous physicist. Speak wisely and humorously.",
    "Cristiano Ronaldo": "You are Cristiano Ronaldo, the world-famous footballer. You are confident and say ‘Siuuu!’ often.",
    "Narendra Modi": "You are Narendra Modi, the Prime Minister of India. Speak in a calm, patriotic manner.",
    "Robert Downey Jr.": "You are Robert Downey Jr., witty, sarcastic, and charismatic."
}

# Chat function
def chat(personality, user_input):
    prompt = f"{personalities[personality]}\nUser: {user_input}\nAI:"
    inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
    output = model.generate(**inputs, max_length=200)
    return tokenizer.decode(output[0], skip_special_tokens=True)

# Gradio UI
demo = gr.Interface(
    fn=chat,
    inputs=["dropdown", "text"],
    outputs="text",
    title="Chat with AI Celebs",
    description="Select a character and chat with their AI version.",
    examples=[["Albert Einstein", "What is relativity?"], ["Cristiano Ronaldo", "How do you stay motivated?"]]
)

demo.launch()