lexn 查看onnx模型结构-使用Netron模块

 

netron模型介绍

https://blog.csdn.net/qq_38806886/article/details/126464547

 点击模型输入input的graph，可以看到使用的torch版本信息，这里是torch1.13,对应onnx是v13，解释一下onnx显示的v13版本是 torch.onnx.export函数中opset_version对应的输出版本。13对应onnxv7，11对应onnxv6

查看onnx模型结构-使用Netron模块

1 安装

$pip install netron

2 可选-查看安装的路径

$ pip show netron

3 新建：python look_onnx.py，复制如下内容，路径修改为本次onnx文件路径

1 2 3 4 5 6 7

import netron   # �??�?ONNX模�??�??件�??路�? onnx_model_path = r'd:/yolo5/yolov5n-seg_toXiaoLiu/model/yolov5n-seg.onnx'   # �?�"�netron�?��?��?�??ONNX模�?? netron.start(onnx_model_path)

　　查看onnx结构，

$python look_onnx.py

4 浏览器中即打开onnx

(base) C:\Users\tt>conda activate pytorch

(pytorch) C:\Users\tt>pip install netron
Looking in indexes: https://mirrors.aliyun.com/pypi/simple/
Collecting netron
  Downloading https://mirrors.aliyun.com/pypi/packages/18/11/d5fa133ef0b7340d4c37022372f3959190e793d2a38231e13a29e7f073e4/netron-8.4.2-py3-none-any.whl (1.9 MB)
     ---------------------------------------- 1.9/1.9 MB 3.6 MB/s eta 0:00:00
Installing collected packages: netron
Successfully installed netron-8.4.2

(pytorch) C:\Users\tt>pip show netron
Name: netron
Version: 8.4.2
Summary: Viewer for neural network, deep learning and machine learning models.
Home-page:
Author:
Author-email: Lutz Roeder <lutzroeder@users.noreply.github.com>
License: MIT
Location: d:\anaconda3\envs\pytorch\lib\site-packages
Requires:
Required-by:

(pytorch) C:\Users\tt>python look_onnx.py
python: can't open file 'look_onnx.py': [Errno 2] No such file or directory

(pytorch) C:\Users\tt>d:

(pytorch) D:\>cd deep_learning

(pytorch) D:\deep_learning>cd yolov5-master

(pytorch) D:\deep_learning\yolov5-master>cd yolov5-master
系统找不到指定的路径。

(pytorch) D:\deep_learning\yolov5-master>python look_onnx.py
Serving 'D:/deep_learning/yolov5-master/best.onnx' at http://localhost:8080