clockwiser

Different numpy versions have different implementations of the + operator. See below link. https://stackoverflow.com/questions/32909619/numpy-array-1-9-2-getting-valueerror-could-not-broadcast-input-array-from-shape

Finally solved the problem, although it is really an ugly way of handling the problem. There should be a better way. train_data = [None] * 13 train_data[0] = z1_train train_data[1]...

Each number has each own z_mean. And we can assume each number has same z_var. The point of vae is trying to put all z_means into Normal(0,1) space.

initializer creates constant tensors, so is the error. Where can find the old version source of initializations.py?

After find old version, changed the custom batch normalization, yet problems there are. Later found keras BatchNormalization is based on the same theory as below. https://arxiv.org/abs/1502.03167 So switched to keras...

Python2 could be the cause of the problem. Try python 3.

frideos 1.0.1 seems to work on web. But it doesn't support null safe. Warnings pop up: The library 'package:frideos/frideos.dart' is legacy, and should not be imported into a null safe...

In the typescript documents, it says each enum member has a value associated with it which can be either constant or computed. https://www.typescriptlang.org/docs/handbook/enums.html enum FileAccess { // constant members None,...