import sys, os
from mixed import igan
from dataset import cook_split_base as base_dataset
if __name__ == '__main__':
# from dataset import cook_mnist_complete as base_dataset
base_splits = { "base" : [0,1,2,3,4,5,6,7,8,9],
"shot" : [],
"p" : 0 }
base = base_dataset (splits = base_splits, verbose = 1)
base = base.dataset_location()
# base = '_datasets/_dataset_63616'
# This will initialize the igan. Both MLP and GAN
# will be training with the base dataset.
root = 'records/site_1'
if not os.path.exists(root):
os.makedirs(root)
# initialize igan object
site1 = igan ( init_dataset = base, root = root, verbose = 1 )
# setup and train site-1 Base MLP
"""site1.setup_base_mlp(root = root, verbose = 2)
lr = (0.04, 0.01, 0.0001)
epochs =(15, 15)
site1.train_base_mlp ( lr =lr,
save_after_epochs = 1,
epochs = epochs,
early_terminate = False,
import sys, os
from svhn import igan
from dataset import cook_split_base as base_dataset
if __name__ == '__main__':
# from dataset import cook_mnist_complete as base_dataset
base_splits = {"base": [0, 1, 2, 3, 4, 5], "shot": [6, 7, 8, 9], "p": 0}
data_loc = '/home/local/ASUAD/rvenka10/Dropbox/svhn/' # this is where you will add the matlab
# data files as processed from the
# yann mat2yann tutorial.
base = base_dataset(location=data_loc, splits=base_splits, verbose=1)
base = base.dataset_location()
# Data was once made and stored here. Uncomment the lines above
# instead of the one below to newly make one.
# base = '_datasets/_dataset_88730'
# This will initialize the igan. Both MLP and GAN
# will be training with the base dataset.
root = 'records/site_1'
if not os.path.exists(root):
os.makedirs(root)
# initialize igan object
site1 = igan(init_dataset=base, root=root, verbose=1)
# setup and train site-1 Base MLP
site1.setup_base_mlp(root=root, verbose=1)
lr = (0.04, 0.001)
epochs = (15)