- Fixed Bug if number of batches is bigger than latent dim - Fixed Bug if only batch in data and mmd loss is on - Better usability of mse loss models
0.3.0
- switch to PyTorch
0.2.2
- get_latent bug fixed - ZINB loss calculation modified - minor improvements
0.2.1
0.1.5
Minor bugs fixed:
- removed tf-gpu 1.15.2 from requirements (not necessary in requirements) - moved pancreas test to development branch - set_random_seed used for networks in master branch (TF < 2)