can't find dataset['train_mix']
wx3771 opened this issue · 3 comments
wx3771 commented
Columbine21 commented
Line 40 in 6eb09e5
train_mix is an augmentation from the original train set. Using the following operations.
wx3771 commented
lyh512796310 commented
Could you please provide this ? Thank you!
For the convenience of downloading the newly provided dataset, we divided the supervised data and unsupervised data into two ".pkl" files, and "train_mix” will put the supervised data and unsupervised data together. You can concatenate supervised and unsupervised data together as "train_mix" data. In addition, "mask" is set to 1 for supervised data and 0 for unsupervised data. For example, the data corresponding to MASK: [1,1,0,0] is Data: [supervised data, supervised data, unsupervised data, unsupervised data].