mini-imagenet and tiny-imagent dataset transformation for traditional classification task and also for the format for few-shot learning / meta-learning tasks
Why do you think that https://github.com/qubvel/segmentation_models.pytorch is a good alternative to MLclf