app.py

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

# 加载模型和分词器
model_name = "taide/Llama3-TAIDE-LX-8B-Chat-Alpha1"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float16, device_map="auto")

# 定义推理函数
def generate_text(prompt):
    inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
    outputs = model.generate(**inputs)
    return tokenizer.decode(outputs[0], skip_special_tokens=True)

# 创建 Gradio 界面
interface = gr.Interface(
    fn=generate_text,
    inputs="text",
    outputs="text",
    title="Llama3 Chatbot",
    description="与 Llama3-TAIDE-LX-8B-Chat-Alpha1 交谈"
)

# 启动 Gradio 应用
interface.launch()