思路:使用模型的特征提取层,转化成向量,然后比对向量的距离(比如cos)
import torch
import torchvision.models as models
import torchvision.transforms as transforms
from PIL import Image
# load model & feature only
model = models.mobilenet_v3_small(pretrained=True)
mode[......]
思路:使用模型的特征提取层,转化成向量,然后比对向量的距离(比如cos)
import torch
import torchvision.models as models
import torchvision.transforms as transforms
from PIL import Image
# load model & feature only
model = models.mobilenet_v3_small(pretrained=True)
mode[......]
import cv2
import numpy as np
import onnxruntime as ort
def load_model(model_path):
"""加载ONNX模型"""
session = ort.InferenceSession(model_path)
return session
def preprocess_image(image_path):
image = cv2.imread(image_path)[......]
#include <opencv2/opencv.hpp>
#include <onnxruntime/onnxruntime_cxx_api.h>
#include <vector>
#include <iostream>
int main() {
// load onnx model
Ort::Env env(OrtLoggingLevel::ORT_LOGGING_LEVEL_WARNING, "test"[......]
之前Docker时有个--restart always的参数,能实现开机自启(原意是挂掉后拉起)
在切换到podman后无效了,原因是podman设计的很轻量,没有守护进程,解决方案如下:
生成systemd文件:
podman generate systemd --name your-container-name --files --new
上述your-container-name是你正在运行的容器,会自动提取对应参数
sudo mv container-your-cont[......]
下载:https://github.com/microsoft/onnxruntime/releases
sudo mkdir /usr/lib/onnxruntime
sudo mkdir /usr/include/onnxruntime
mv path/onnxruntime-linux-x64-1.17.3/lib/* /usr/lib/onnxruntime
mv path/onnxruntime-linux-x64-1.17.3/include/* /usr/include/o[......]