Gregory1994
Gregory1994
作者有提供模型哦。。。你可以看一下这个Issue #2 使用方法的话,可以试试这样: ``` python predict.py --test-path ./image ```
Of course you can't run train.py without GPUs...
similar problem, any solution please? 
You can try to change the threshold in './data/config.py'. for example, you can change "widerface_640['conf_thresh'] = 0.6" to "widerface_640['conf_thresh'] = 0.1"