How can resume the train_search.py using weights in log_dir if the code stops due to accidents like power cut ... ?

chenxin061 / pdarts

Codes for our paper "Progressive Differentiable Architecture Search:Bridging the Depth Gap between Search and Evaluation"

Other

360 stars 83 forks source link

Closed JarveeLee closed 5 years ago

chenxin061 commented 5 years ago

You can resue it with regular PyTorch functions such as torch.load() and model.load_state_dict().