The program stops automatically

[wanzhixiao]

When run this program, the gpu memory used is always changing. After running several epochs, it stops automatically.

[wanzhixiao]

When run this program, the gpu memory used is always changing. After running several epochs, it stops automatically.

I have found the problem,it is this line that cause problem. in train.py i comment out this line of code ,torch.cuda.empty_cache(),it works. the reason why CPU memory increases while training is that, dataloader set too many workers.

[shuqincao]

你好，请问你怎么把代码跑起来的呢

[wanzhixiao]

你好，请问你怎么把代码跑起来的呢

你好，我是用其他的数据集，根据config目录下evoconv2-config.yaml文件中的设置，修改了数据加载中的代码。

[shuqincao]

你好，请问能加你微信吗？1033898863

[eeluo]

你好，请问你怎么把代码跑起来的呢

你好，我是用其他的数据集，根据config目录下evoconv2-config.yaml文件中的设置，修改了数据加载中的代码。

请问您成功运行了吗？我怎么总是遇到矩阵的维数有问题？可以留个联系方式吗？谢谢