search

dog-qiuqiu / Yolo-Fastest

:zap: Based on yolo's ultra-lightweight universal target detection algorithm, the calculation amount is only 250mflops, the ncnn model size is only 666kb, the Raspberry Pi 3b can run up to 15fps+, and the mobile terminal can run up to 178fps+
Other
1.98k stars 427 forks source link

请问如果我想训练自己的数据，需要怎么更改.cfg文件？ #17

Closed JimXu1989 closed 3 years ago

JimXu1989 commented 3 years ago

您好！

如果我想训练自己的数据集，需要如何更改.cfg文件？我看这个项目中的.cfg文件格式和darknet yolov4的还是不太一样，比如我的classes数量和预训练的模型不一样，我需要按什么规则更改呢？

QAbot-zh commented 3 years ago

这个cfg是类似yolov3的配置，修改classes， 然后每个[yolo]层的前一层卷积核数量（filters）改为（classes+1）* 3

zxj11838 commented 3 years ago

You can refer to this part of （How to train ）the project（https://github.com/AlexeyAB/darknet）.

Lebhoryi commented 3 years ago

You can refer to this part of （How to train ）the project（https://github.com/AlexeyAB/darknet）.

sry, the link is invalid now

zxj11838 commented 3 years ago

@Lebhoryi Copy the URL and enter it into the browser.

Lebhoryi commented 3 years ago

@zxj11838 It works now.

There is a new question, how could I convert model to '.h5' model successfully?

zxj11838 commented 3 years ago

Sorry, I have not converted.