I've tried many ways to prevent nan (normalizing images, low lr) , but it still persists. I've tried with two different datasets. Both of them give nan loss in first epoch. Only 'loss_rpn_box_reg' becomes nan. What can be the reason ? I also discussed this issuehere.Another thread...
百度试题 结果1 题目RPN中reg loss,即rpn_loss_bbox层计算的softmax loss,用于bounding box regression网络训练。 A. 正确 B. 错误 相关知识点: 试题来源: 解析 B 反馈 收藏
RPN中reg loss,即rpn_loss_bbox层计算的softmax loss,用于bounding box regression网络训练 A. 正确 B. 错误 题目标签:网络训练计算如何将EXCEL生成题库手机刷题 如何制作自己的在线小题库 > 手机使用 分享 反馈 收藏 举报 参考答案: B 复制 纠错