Skip to main content

Advertisement

Table 1 Hyperparameter settings used for CVAE training

From: Deep clustering of protein folding simulations

Hyperparameter Lower bound Upper bound Optimal Fs-Peptide Optimal BBA Optimal VHP
Num convolutional layers 1 4 4 4 4
Num convolutional filters 16 125 100 100 99
Convolutional filter shape (m×m) 2 7 5 5 5
Num dense neurons 32 100 64 64 58
Latent dimension 2 16 10 10 9
Mean reconstruction error    7.82 24.31 74.66
Mean reconstruction error (latent dim 3)    23.24 53.48 113.07