app.py

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

# 加载 DeepSeek 聊天模型
model_name = "deepseek-ai/deepseek-llm-67b-chat"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")

def chat_with_deepseek(user_input):
    messages = [{"role": "user", "content": user_input}]
    input_tensor = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt")
    outputs = model.generate(input_tensor.to(model.device), max_new_tokens=100)
    result = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return result

# 创建 Gradio 接口
demo = gr.Interface(fn=chat_with_deepseek, 
                    inputs="text", 
                    outputs="text", 
                    title="DeepSeek Chatbot",
                    description="与 DeepSeek 聊天机器人对话！")

# 启动应用程序
if __name__ == "__main__":
    demo.launch()