diff options
| author | David Luevano <55825613+luevano@users.noreply.github.com> | 2019-12-28 11:12:36 -0700 |
|---|---|---|
| committer | David Luevano <55825613+luevano@users.noreply.github.com> | 2019-12-28 11:12:36 -0700 |
| commit | e4f9e15588ec796f73c000a683cc9152454a913c (patch) | |
| tree | 698f24d36cbf851e263625adb883339efaaacdce | |
| parent | c1e7b327655ebaa5c44e4bef5b9b675b23782952 (diff) | |
Fix bugs
| -rw-r--r-- | lj_matrix/__main__.py | 10 | ||||
| -rw-r--r-- | lj_matrix/do_ml.py | 7 |
2 files changed, 14 insertions, 3 deletions
diff --git a/lj_matrix/__main__.py b/lj_matrix/__main__.py index 98f341e1e..811024ff0 100644 --- a/lj_matrix/__main__.py +++ b/lj_matrix/__main__.py @@ -24,6 +24,14 @@ from lj_matrix.do_ml import do_ml # from lj_matrix.misc import plot_benchmarks if __name__ == '__main__': - do_ml(min_training_size=1500, max_training_size=3000) + do_ml(min_training_size=1500, + max_training_size=2000, + training_increment_size=500, + test_size=None, + ljm_diag_value=None, + ljm_sigma=1.0, + ljm_epsilon=1.0, + save_benchmarks=False, + show_msgs=True) # plot_benchmarks() print('OK!') diff --git a/lj_matrix/do_ml.py b/lj_matrix/do_ml.py index 45dc7a5f0..da9386bf7 100644 --- a/lj_matrix/do_ml.py +++ b/lj_matrix/do_ml.py @@ -114,6 +114,7 @@ def ml(desc_data, def do_ml(min_training_size, max_training_size=None, training_increment_size=500, + test_size=None, ljm_diag_value=None, ljm_sigma=1.0, ljm_epsilon=1.0, @@ -128,6 +129,8 @@ def do_ml(min_training_size, min_training_size: minimum training size. max_training_size: maximum training size. training_increment_size: training increment size. + test_size: size of the test set to use. If no size is given, + the last remaining molecules are used. ljm_diag_value: if a special diagonal value should be used in lj matrix. ljm_sigma: sigma value for lj matrix. ljm_epsilon: epsilon value for lj matrix. @@ -170,7 +173,7 @@ def do_ml(min_training_size, i, 'CM', cm_send, - max_training_size, + test_size, sigma, show_msgs)) procs.append(p1) @@ -184,7 +187,7 @@ def do_ml(min_training_size, i, 'L-JM', ljm_send, - max_training_size, + test_size, sigma, show_msgs)) procs.append(p2) |