Files are organized as follows:
- Preprocessing of text:
preprocess. - Pretraining of text:
transformer_encoder. - Pretraining of board:
katago. - Test cases or examples for pretarined text and board:
test. - Baseline:
logistic_regression. - NN model without pretrained features:
nn_model. - Combining pretrained text and board embeddings:
pretrain_combine.