Use Custom Training Data


Can we create custom training datasets to train the CVAE? Currently there are 100 x0 & x1 vectors with their associated given_y, which are used for training. Are we allowed to create & use custom datasets to improve the training process?


Hello @johnabish00
Thank you for your question.

Regrettably, you are limited to using only datasets provided by the Xeek Team.

Please note that the provided training dataset consists of 90,000 pairs of x0 and x1 with corresponding given_y values, not 100. The batch size in the Starter notebook is set to 100 for convenience, but you are free to adjust it according to your needs and comprehension.

Xeek Team