resnet43_local_34.ini

[info]
experiment: ResNet default settings
name: V2_resnet43_34

[dataset]
pixels: 64
n_classes: 2
channels: 1
subset: None
filenames_train: ./../../data/cadV2_0.5mm_64x64_xy_xz_yz/subset[0152789]/*/*.pkl.gz
filenames_validation: ./../../data/cadV2_0.5mm_64x64_xy_xz_yz/subset[6]/*/*.pkl.gz
data_folder: None ;not used

[network]
architecture: resnet
input_size: 64
depth: 4
branching_factor: 3
batch_normalization: True
batch_normalization_alpha: 0.1
dropout: 0.25
spatial_dropout: 0.5
gaussian_noise: 0.05

[updates]
optimization: -
learning_rate: 0 ;not used, see schedule
momentum: 0.92
l2_lambda: 0.00003
batch_size_train: 60
batch_size_validation: 140
n_epochs: 240

[preprocessing] ;Not used
random_crop: 0 ;no cropping
erode_segmentation: 11

[normalization]
zero_center: True
mean_pixel: 0.2606236106512

[augmentation]
augment: True
flip: True
zoom: 0.08
rotation: 20
translation: 3


[misc]
multiprocess_load_augmentation: False
save_every_n_epoch: 2
n_workers_load_augmentation: 4