Repository for various machine learning implementations for me research.
Given a very specific setup of training data:
- .pkl files are all located in a separate directory:
prep-pkl
- Update the location of the desired .pkl file in
learnme.py
(or relevant notebook) - Run:
./learnme.py
with the following optional arguments (they indicate True if present, False if not): 1.--track_preds
(-tp
), 2.--err_n_scores
(-es
), 3.--learn_curves
(-lc
), 4.--valid_curves
(-vc
), and 5.--test_compare
(-tc
) - cleanup and store results; update results paths in plotme.py
- Graph results via:
./plotme.py