jonasrothfuss / meta_learning_pacoh

Meta-learning Gaussian process (GP) priors via PAC-Bayes bounds
MIT License
25 stars 13 forks source link

Fix bug where MNIST train and test are the same #5

Open christabella opened 4 years ago

christabella commented 4 years ago

But maybe this doesn't matter, since both the meta train and meta test data are from generate_meta_test_data? Is the reason for this that for the synthetically generated datasets it doesn't matter?

https://github.com/jonasrothfuss/meta_learning_pacoh/blob/7098ca212c24dd73054bc32e2fa6050e07fbdc4b/experiments/meta_overfitting_v2/meta_GPR_overfitting_base.py#L75-L79