Zhonghan Deng
Zhonghan Deng
you did not set cuda properly. the training is running on cpu. This might be the reason
你这明显有问题啊, 怎么都不可能这么差啊
我训练都挺正常的, 就是训练到最后的mAP不如YOLOV5. 你可以可视化一下你的label, 也可以检查一下你的label有没有跟图片尺寸normalize.
配置文件用的都是default的, 没改变.
那你训练后面几百个epoch也是这么低m吗
Thanks for your reply. What about inferencing? (infer.py)
Hello, thanks for your reply. I am not sure if I understand it correctly or not. When using infer.py to infer,  It does not have rbr_reparam. I did check...
Will do, thanks.
This error means your img_info is empty. You can just debug this. It is easy to fix.
Some other implementation for yolov3 would find the best anchor for each true box. Is there any gap I am having? I can understand that each true box would be...