CentOS PyTorch模型部署与优化策略

sudo yum install python3 python3-pip
pip3 install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu113

pip3 install torch torchvision torchaudio

from flask import Flask, request, jsonify
import torch
from your_model import YourModel  # 假设你的模型定义在这个模块中

app = Flask(__name__)
model = YourModel()
model.load_state_dict(torch.load('your_model.pth'))
model.eval()

@app.route(<span><span>'/predict', methods=['POST']</span>)</span>
def predict():
    data = request.json['data']  # 假设data是一个包含模型输入的张量数据
    input_tensor = torch.tensor(data['input'])
    with torch.no_grad():
        output = model(input_tensor)
    return jsonify(output.tolist())

if __name__ == '__main__':
    app.run(host='0.0.0.0', port=5000)

nohup python3 your_service_script.py &

FROM python:3.8 -slim

# 安装PyTorch和其他依赖
RUN pip install torch torchvision torchaudio

# 复制模型和服务代码到容器中
COPY your_model.pth /app/your_model.pth
COPY your_service_script.py /app/your_service_script.py

# 设置工作目录
WORKDIR /app

# 运行服务
CMD [<span>"python", "./your_service_script.py"]</span>

docker build -t your-model-service .
docker run -p 5000:5000 your-model-service