config.yaml

neuroscope:
  source      : img         # fmri or img
  image_size  : 64          # only applicable to COCO (neuroscope) images
  latent_dim  : 128          # bottleneck size (latent space)
  batch_size  : 64
  in_chans    : 1           # number of input channels
  chan_start  : 32           # number of output channels for first conv layer
  conv_branch : 2           # number of branches for conv layers
  kernel_size : 3           # 3 or 5
  stride      : 2           # 1 or 2 or 4
  gcn_layers  : 4           # 2 or 3
  conv_layers : 5           # could be quite deep i think
  fc_layers   : 1           # should probably be 1
  lr          : 0.001       # seems 0.001 is best even with batch norm
  epochs      : 50          # between 10 and 50
  batch_norm  : True        # True or False
  vae         : False       # True or False
  beta        : 0.5         # beta for beta-VAE
  loss_fn     : l1          # l1 or l2
  dropout     : 0.0       # dropout rate for fc layers