FoldingNet Model

Import from molearn.models.foldingnet

class AutoEncoder(*args, **kwargs)

Autoencoder architecture derived from FoldingNet.

Initializes internal Module state, shared by both nn.Module and ScriptModule.

CNN model

import from molearn.models.CNN_autoencoder

class Autoencoder(init_z=32, latent_z=1, depth=4, m=1.5, r=0, droprate=None)

This is the autoencoder used in our Ramaswamy 2021 paper. It is largely superseded by molearn.models.foldingnet.AutoEncoder().

Parameters:

• init_z (int) – number of channels in first layer

• latent_z (int) – number of latent space dimensions

• depth (int) – number of layers

• m (float) – scaling factor, dictating number of channels in subsequent layers

• r (int) – number of residual blocks between layers

• droprate (float) – dropout rate