为何在ICDAR2015训练 val_loss0.73就再也下不去了？

[lijian10086]

使用默认参数训练736的模型，ICDAR2015的1000张图片，train=9:val=1,val_loss很快过拟合，版主有遇到过么？如何解决？@huoyijie

[lijian10086]

train-loss=0.41

[Zhang-O]

大哥 你是拿着预训练模型 接着训练的吗

[isyanan1024]

@lijian10086 请问训练的时候cuda和cudnn版本是多少，一直提示 Loaded runtime CuDNN library: 7600 (compatibility version 7600) but source was compiled with 7102 (compatibility version 7100). If using a binary install, upgrade your CuDNN library to match. If building from sources, make sure the library loaded at runtime matches a compatible version specified during compile configuration.

[wsy915]

@lijian10086 请问你的这个问题解决了吗，我用主页给的这10000个样本的数据集跑，也出现这个问题。在epoch=12就停止了

[v-smwang]

+1

[xiulianzw]

@lijian10086 请问训练的时候cuda和cudnn版本是多少，一直提示 Loaded runtime CuDNN library: 7600 (compatibility version 7600) but source was compiled with 7102 (compatibility version 7100). If using a binary install, upgrade your CuDNN library to match. If building from sources, make sure the library loaded at runtime matches a compatible version specified during compile configuration.

你升级一下tensorflow-gpu的版本，装个1.8的试试