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

# Load model and tokenizer
model_name = "DAMO-NLP-SG/VideoLLaMA3-7B"
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True)
model.eval()

def generate_response(prompt, max_tokens=200, temperature=0.7):
    inputs = tokenizer(prompt, return_tensors="pt")
    with torch.no_grad():
        output = model.generate(
            **inputs,
            max_new_tokens=max_tokens,
            temperature=temperature,
            do_sample=True,
            top_p=0.9,
            eos_token_id=tokenizer.eos_token_id
        )
    response = tokenizer.decode(output[0], skip_special_tokens=True)
    return response[len(prompt):].strip()  # Return only the generated part

# Gradio UI
iface = gr.Interface(
    fn=generate_response,
    inputs=[
        gr.Textbox(label="Prompt", lines=5, placeholder="Enter your prompt here..."),
        gr.Slider(minimum=50, maximum=1000, value=200, label="Max Tokens"),
        gr.Slider(minimum=0.1, maximum=1.0, value=0.7, label="Temperature"),
    ],
    outputs=gr.Textbox(label="Response"),
    title="VideoLLaMA3-7B Text Generation",
    description="Generate text using DAMO-NLP-SG/VideoLLaMA3-7B"
)

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