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

# Read Hugging Face token from environment variable
HF_TOKEN = os.getenv("HF_TOKEN")


MODEL_NAME = "mistralai/Mistral-7B-v0.1"

print("Loading tokenizer and model...")
tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, use_auth_token=HF_TOKEN)
model = AutoModelForCausalLM.from_pretrained(MODEL_NAME, use_auth_token=HF_TOKEN)

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

def generate_text(prompt):
    inputs = tokenizer(prompt, return_tensors="pt").to(device)
    outputs = model.generate(
        **inputs,
        max_length=150,
        do_sample=True,
        temperature=0.7,
        top_p=0.9,
        eos_token_id=tokenizer.eos_token_id
    )
    text = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return text

# Gradio UI
iface = gr.Interface(
    fn=generate_text,
    inputs=gr.Textbox(lines=5, placeholder="Enter your prompt here..."),
    outputs="text",
    title="TechChat - Mistral 7B",
    description="Generate text with hari7261/TechChat model hosted on Hugging Face."
)

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