Training in Section 4 To train ProtoNet with Resnet-12 backbone using FixML objective on mini-imagenet, use the following scripts:

bash scripts/resnet-12/miniImagenet/train_5w5s_fixS_PN.sh

To train ProtoNet with Resnet-12 backbone using ML objective, use the following scripts:

bash scripts/resnet-12/miniImagenet/train_5w5s_metal_PN.sh

Evaluations in Section 4.2, 5.1 The evaluation script is in eval.py which would evaluate a given algorithm over a sequence of task distributions between base and novel.

Evaluations in Section 5.2, Appendix D The variance analysis is done using python scripts in analysis/compute_novel_acc_variance.py and analysis/compute_base_acc_variance.py.