Comet
무료로 딥러닝 실험 결과 및 training 상황을 저장하고, 비교하고 관리할 수 있게 해 주는 툴.
Installation
mamba install -c comet_ml comet_ml
Shell
복사
Define experiment
from comet_ml import Experiment
experiment = Experiment(
api_key='XAHbm......',
project_name='my_project',
workspace='dohlee',
)
Python
복사
log_parameters
hyperparameters = {
'learning_rate': 0.05,
'batch_size': 128,
}
experiment.log_parameters(hyperparameters)
Python
복사
log_metric
experiment.log_metric('train_loss', running_loss / show_every, step=step)
Shell
복사
log_histogram_3d
experiment.log_histogram_3d(net.state_dict()['fc.9.weight'].cpu(), step=step)
Python
복사