我遇到了一个问题,我在自己的数据集上训练了py-faster-rcnn,但它在迭代660中长时间停留并且没有响应。以前有人遇到过这个问题吗?怎么解决?非常感谢!
FYR:
I0404 17:50:23.514286 25601 solver.cpp:571] Iteration 640, lr = 0.001 I0404 17:50:24.876749 25601 solver.cpp:242] Iteration 660, loss = 0.341694 I0404 17:50:24.876804 25601 solver.cpp:258] Train net output #0: rpn_cls_loss = 0.296529 (* 1 = 0.296529 loss) I0404 17:50:24.876819 25601 solver.cpp:258] Train net output #1: rpn_loss_bbox = 0.0451643 (* 1 = 0.0451643 loss) I0404 17:50:24.876835 25601 solver.cpp:571] Iteration 660, lr = 0.001