batch_cutoff not working for split = 'train'

katerakelly / pytorch-maml

PyTorch implementation of MAML: https://arxiv.org/abs/1703.03400

MIT License

553 stars 129 forks source link

batch_cutoff not working for split = 'train' #16

Open pppplin opened 5 years ago

pppplin commented 5 years ago

Hi Kate,

Under split='train' in inner loop, the data loader always output the same data no matter under what batch_cutoff. I know that this is the setting in the MAML paper, but I am wondering is there a way to make batch_cutoff work? Many thanks!

pppplin commented 5 years ago

think this is related to #9