团长CXYsama

System: win10 1903 Velocity:1.1.14 Vscode:1.36 Vscode:2.2.0

> this error is still exits in 2020/4/29. >  use `https://keras-zh.readthedocs.io/` is OK

= =||| I try add ```bash " When Tmux 'focus-events' option is on, Tmux will send [O when the " window loses focus and [I when it gains focus. exec...

I use the `cmake-gui` to configure the `VLFEAT_LIBRARY` successfully. firstly: ```bash cd ./vlfeat-0.9.20/ make ``` If make success,in your `vlfeat-0.9.20` dir you can find `bin` folder the just run `cmake-gui`...

@dgl547437235 你好，我想请教以下您这里用的最大编码距离是什么距离？ 因为代码的test里写的是用欧式距离，但是论文里分数用的L1范数。。。我将0设置为alnomaly，在mnist上训练了15轮之后，测试的时候发现NetG对0的重建效果也很好= =|||，分数无论用L1范数还是L2范数似乎都不能很好的找个阈值来区分正常和异常样本。

meet same issue . OS: `Ubuntu 18.04` vim version: ``` NVIM v0.5.0-dev+1157-g0ab88c2ea Build type: Release LuaJIT 2.1.0-beta3 ``` Ctags version: ``` Universal Ctags 5.9.0(p5.9.20201206.0), Copyright (C) 2015 Universal Ctags Team...

do you use tmux?I meet same question,here is some idea that maybe can fix it:https://github.com/tmux-plugins/vim-tmux-focus-events/issues/6

I think 1280 >720,so model will padding 720 to 736 instead of simple resize...The effect of this operation on the performance of the model should be small

maybe you analyze label file though mutil thread,then use add_sample manualy... I use this method replace use Importer, sample import speed increased by 5 times... here is a glance of...

[transforms.ToTensor](https://pytorch.org/vision/stable/transforms.html#torchvision.transforms.ToTensor) will norm image to [0,1]....