Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

训练YOLOX-S #45

Closed
vaerdu opened this issue Sep 3, 2022 · 8 comments
Closed

训练YOLOX-S #45

vaerdu opened this issue Sep 3, 2022 · 8 comments

Comments

@vaerdu
Copy link

vaerdu commented Sep 3, 2022

用自己的数据集训练本仓库的YOLOX-S比官方YOLOX-S要高6-7个点,本仓库YOLOX代码和官方YOLOX代码哪里不一样吗?

@iscyy
Copy link
Owner

iscyy commented Sep 4, 2022

Compared with the official, there are some optimizations, and the training method adopts the parameter configuration of yolov5

@vaerdu
Copy link
Author

vaerdu commented Sep 4, 2022

Compared with the official, there are some optimizations, and the training method adopts the parameter configuration of yolov5

对比了下,本仓库的YOLOX-S参数量约为8.0M,官方YOLOX-S参数量是8.9M,请问是改变了什么模块吗?只是超参数的调整不应该参数量不一致吧

@iscyy iscyy closed this as completed Sep 4, 2022
@lunaWU628
Copy link

用自己的数据集训练本仓库的YOLOX-S比官方YOLOX-S要高6-7个点,本仓库YOLOX代码和官方YOLOX代码哪里不一样吗?

你好,请问你是怎么训练的yolox呢?这个存储库不是yolov5的嘛

@vaerdu
Copy link
Author

vaerdu commented Sep 19, 2022

用自己的数据集训练本仓库的YOLOX-S比官方YOLOX-S要高6-7个点,本仓库YOLOX代码和官方YOLOX代码哪里不一样吗?

你好,请问你是怎么训练的yolox呢?这个存储库不是yolov5的嘛
仓库里,configs文件夹里有yolox的网络结构

@lunaWU628
Copy link

嗯嗯,看到了,所以我想用这个存储库的yolox进行训练的话,把这个命令:
python path/to/train.py --data coco128.yaml --weights '' --cfg yolov5s.yaml --img 640
改成:
python path/to/train.py --data coco128.yaml --weights '' --cfg yolox.yaml --img 640 就可以了嘛

@vaerdu
Copy link
Author

vaerdu commented Sep 19, 2022

嗯嗯,看到了,所以我想用这个存储库的yolox进行训练的话,把这个命令: python path/to/train.py --data coco128.yaml --weights '' --cfg yolov5s.yaml --img 640 改成: python path/to/train.py --data coco128.yaml --weights '' --cfg yolox.yaml --img 640 就可以了嘛

中间可能会报一些错误,根据错误修改就跑通了,在issue可以查找解决办法

@lunaWU628
Copy link

好的,感谢~

@sneepbob
Copy link

请问,在训练yolox的时候,训练到一定epoch时,发生报错:

assertion input_val zero && input_val one cuda trigger
RuntimeError: CUDA error: device-side assert triggered

怎么解决呢?
感谢大佬们解答!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants