You can configure the parameters to be adjusted in the config.py file, or simply embed trainer.py into your project. In config.py you need to configure the following settings:
seed = 2022108
epochs = 30
batch_size = 64
learning_rate = 0.003
resume_path = None
fix_params = {
'threshold':0.9,
'mu':7,
'lambda':1,
'T':1
}
# labeled data num
## dataset ['mnist','cifar10','stl10','pets']
eval_iters = 100
dataset = "mnist"
num_labeled = 100
root_path = r'./data'
num_classes = 5
run
python main.py