app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer
from vllm import LLMEngine, SamplingParams

# Load the model and tokenizer from Hugging Face
model_name = "Qwen/Qwen2-7B"  
tokenizer = AutoTokenizer.from_pretrained(model_name)
engine = LLMEngine(model=model_name)

def generate_response(prompt, max_tokens, temperature, top_p):
    # Tokenize the prompt
    inputs = tokenizer(prompt, return_tensors="pt")
    
    # Define sampling parameters
    sampling_params = SamplingParams(
        max_tokens=max_tokens,
        temperature=temperature,
        top_p=top_p,
    )
    
    # Generate text using vLLM
    output = engine.generate(inputs["input_ids"], sampling_params)
    
    # Decode the generated tokens to text
    generated_text = tokenizer.decode(output[0]["token_ids"], skip_special_tokens=True)
    return generated_text

# Gradio UI
with gr.Blocks() as demo:
    gr.Markdown("# 🚀 Hugging Face Integration with vLLM")
    gr.Markdown("Generate text using the vLLM integration with Hugging Face models.")

    with gr.Row():
        with gr.Column():
            prompt_input = gr.Textbox(
                label="Prompt",
                placeholder="Enter your prompt here...",
                lines=3,
            )
            max_tokens = gr.Slider(
                label="Max Tokens",
                minimum=10,
                maximum=500,
                value=100,
                step=10,
            )
            temperature = gr.Slider(
                label="Temperature",
                minimum=0.1,
                maximum=1.0,
                value=0.7,
                step=0.1,
            )
            top_p = gr.Slider(
                label="Top P",
                minimum=0.1,
                maximum=1.0,
                value=0.9,
                step=0.1,
            )
            submit_button = gr.Button("Generate")
        
        with gr.Column():
            output_text = gr.Textbox(
                label="Generated Text",
                lines=10,
                interactive=False,
            )
    
    submit_button.click(
        generate_response,
        inputs=[prompt_input, max_tokens, temperature, top_p],
        outputs=output_text,
    )

# Launch the app
demo.launch()