Closed xdg2016 closed 4 years ago
于老师您好,我在使用您的caffe版本训练时,batch_size设置为16,显卡GTX 1060,但是训练起来显存几乎不怎么占用,感觉都用不到gpu的加速,和纯cpu速度差不多,请问有没有什么好的解决办法呢?
模型太小导致,本来计算量小,所以GPU发挥不出威力
于老师您好,我在使用您的caffe版本训练时,batch_size设置为16,显卡GTX 1060,但是训练起来显存几乎不怎么占用,感觉都用不到gpu的加速,和纯cpu速度差不多,请问有没有什么好的解决办法呢?