Open youngsend opened 5 years ago
Here kl_loss formulation uses -5e-4 K.mean(…). However, according to the essay, it should be -0.5 K.sum(…). Here are two implementations: 1. from Keras website: https://keras.io/examples/variational_autoencoder/ from PyTorch examples: https://github.com/pytorch/examples/blob/master/vae/main.py
Here kl_loss formulation uses -5e-4 K.mean(…). However, according to the essay, it should be -0.5 K.sum(…). Here are two implementations: 1. from Keras website: https://keras.io/examples/variational_autoencoder/ from PyTorch examples: https://github.com/pytorch/examples/blob/master/vae/main.py