Esperanto-mega

> Now, I skip the data containing "NaN", and the loss keep 0.135 in epoch 2. > > ``` > [2, 1626] loss: 0.135 Time: 1.461592197418213 sec > [2, 1627]...

raw: assert n_sp_extracted == np.max(superpixels) + 1, ('superpixel indices', np.unique(superpixels)) modified: assert n_sp_extracted == np.max(superpixels), ('superpixel indices', np.unique(superpixels))

> 一般无故卡死是因为多线程读取数据导致内存爆炸，所以大内存很重要，否则就调小num_works 你好 请问这种情况导致的内存爆炸能检测到吗 比如使用命令 watch free -h 还是说卡死之后内存会自动被系统释放