pytorch에서 자주 쓰이는 형태인 numpy, tensor, list 간에 변환. import torch import numpy as np # numpy to tensor np_arr = np.zeros((3, 3), dtype=np.float32) tensor_arr = torch.from_numpy(np_arr) # tensor to numpy np_arr = tensor_arr.numpy() # tensor to list list = tensor_arr.tolist() # list to tensor tensor_arr = torch.tensor(list) 간단하다.

pytorch에서 여러개의 GPU 사용하기. import torch model = MyModel() # CNN이든 뭐든 사용할 모델 device = 'cuda' if torch.cuda.is_available() else 'cpu' if (device == 'cuda') and (torch.cuda.device_count() > 1): model = nn.DataParallel(model) model.to(device) 간단하다.