USE_CUDA = False
if torch.cuda.is_available():
USE_CUDA = True
print('\n\nUSE_CUDA = {}\n\n'.format(USE_CUDA))
img_size = 128
radial_lines = 44
Params_dict = {
'img_size':
img_size, # length of image
'batchsize':
1, # number of samples in batch
'grad_steps':
1, # number of concatenated gradients
'train_steps':
1, # number of optimization steps
'theta':
0.005, # weighting of regularizer
'mask':
GetKMask.createkSpaceMask(np.array([img_size, img_size]),
radial_lines), # mask of radial lines
'optimizer_net':
Architectures.UNet, # optimizer network architecture
'load_model':
True # whether to start new or resume from last saved point
}
solver = Eval.Learnable_Solver(Params_dict)
solver.run()
