Closed xqyd closed 4 years ago
Those pickle-style datasets are obtained from the reference repos, I didn't generate them by meself. Using pickle
package to check their format and store your own dataset with the same format should work.
For using your own dataset, it is not necessary to follow the pickle format (e.g. ImageNet-800 uses image folders), you can write your own pytorch data loader. This may help: https://github.com/cyvius96/few-shot-meta-baseline/issues/5
After writing my own code to split and load the corresponding data into base and novel set respectively, I was able to train and test the code successfully. Thanks for your wonderful work!!!
You're welcome.
Can you share the code or script on how to prepare a given dataset into your pickle-style dataset? Thanks a lot!!!