Hi, I have a huge 1-dimension sparse array (e.g. [0,0,…,0,1,0,0,…,0]) to feed into PyTorch for training. Is there an efficient way to define the input?
I have looked at the torch.sparse, but that seems is for 2-D.
Besides, the targets are also sparse array, is there an efficient way to compare the output of the training and the target?