Fix train/test submodules to use the new test sets

chengsoonong / crowdastro

Cross-identification of radio objects and host galaxies by applying machine learning on crowdsourced training labels.

MIT License

13 stars 1 forks source link

Open MatthewJA opened 8 years ago

MatthewJA commented 8 years ago

They are currently broken since they still rely on is_ir_train etc. from the old version of training data.