app.py

import gradio as gr
import time
from optimum.intel import OVModelForCausalLM
from transformers import AutoTokenizer, pipeline

# Load the model and tokenizer
model_id = "hsuwill000/DeepSeek-R1-Distill-Qwen-1.5B-openvino"
model = OVModelForCausalLM.from_pretrained(model_id, device="CPU")  # 明确指定设备
tokenizer = AutoTokenizer.from_pretrained(model_id)

# Create generation pipeline
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

def respond(message):
    try:
        # Record the start time
        start_time = time.time()
        
        # Only use the current message as input (no history)
        input_text = f"User: {message}"

        # Generate response
        response = pipe(
            input_text,
            max_length=1024,
            truncation=True,
            num_return_sequences=1,
            temperature=0.7,  # 控制生成多样性
            top_p=0.9,        # 控制生成质量
        )
        reply = response[0]['generated_text'].strip()
        
        # Calculate inference time
        inference_time = time.time() - start_time
        print(f"Inference time: {inference_time:.4f} seconds")
        
        # Return as a tuple (user message, bot reply)
        return [(message, reply)]
    
    except Exception as e:
        print(f"Error: {e}")
        return [(message, "Sorry, something went wrong. Please try again.")]

# Set up Gradio chat interface
with gr.Blocks() as demo:
    gr.Markdown("# DeepSeek-R1-Distill-Qwen-1.5B-openvino Chat")
    gr.Markdown("Chat with DeepSeek-R1-Distill-Qwen-1.5B-openvino model.")
    
    chatbot = gr.Chatbot()
    msg = gr.Textbox(label="Your Message")
    
    msg.submit(respond, msg, chatbot)

if __name__ == "__main__":
    demo.launch(share=True)