Merge branch 'uf3:master' into drop_columns_2

monk-04 · Oct 11, 2023 · d1cd6f2 · d1cd6f2
2 parents b1e8dee + 311ade2
commit d1cd6f2
Show file tree

Hide file tree

Showing 6 changed files with 27 additions and 13 deletions.
diff --git a/.gitignore b/.gitignore
@@ -56,3 +56,12 @@ uf3/util/__pycache__/plotting.cpython-37.pyc
 uf3/util/__pycache__/subsample.cpython-37.pyc
 
 lammps_plugin/ML-UF3/.vscode/
+
+.vscode/
+
+*/__pycache__
+
+*/*/__pycache__
+
+*/*/*/__pycache__
+
diff --git a/lammps_plugin/ML-UF3/uf3_pair_bspline.cpp b/lammps_plugin/ML-UF3/uf3_pair_bspline.cpp
@@ -106,4 +106,6 @@ double uf3_pair_bspline::memory_usage()
     bytes += (double)dnbspline_bases[i].memory_usage();     //bspline_basis2
 
   bytes += (double)2*sizeof(double);    //ret_val
+
+  return bytes;
 }
diff --git a/uf3/default_options.yaml b/uf3/default_options.yaml
@@ -48,5 +48,7 @@ learning:
     weight: 0.5
     regularizer:
         ridge_1b: 1.0e-8
+        ridge_2b: 0.0
+        ridge_3b: 1.0e-5
         curvature_2b: 1.0e-8
         curvature_3b: 1.0e-8
diff --git a/uf3/regression/least_squares.py b/uf3/regression/least_squares.py
@@ -499,9 +499,9 @@ def batched_predict(self,
 
         Returns:
             y_e (np.ndarray): target values for energies.
-            p_e (np.ndarray): prediction values for forces.
-            y_f (np.ndarray): target values for energies.
-            p_f (np.ndarray): target values for forces.
+            p_e (np.ndarray): prediction values for energies.
+            y_f (np.ndarray): target values for forces.
+            p_f (np.ndarray): prediction values for forces.
             rmse_e (np.ndarray): RMSE across energy predictions.
             rmse_e (np.ndarray): RMSE across force predictions.
             drop_columns (list): list of columns to drop. Used when modifying
@@ -949,9 +949,9 @@ def subset_prediction(df: pd.DataFrame,
 
     Returns:
         y_e (np.ndarray): target values for energies.
-        p_e (np.ndarray): prediction values for forces.
-        y_f (np.ndarray): target values for energies.
-        p_f (np.ndarray): target values for forces.
+        p_e (np.ndarray): prediction values for energies.
+        y_f (np.ndarray): target values for forces.
+        p_f (np.ndarray): prediction values for forces.
     """
     if subset_keys is not None:
         idx = df.index.unique(level=0).intersection(subset_keys)
@@ -987,9 +987,9 @@ def batched_prediction(model: WeightedLinearModel,
 
     Returns:
         y_e (np.ndarray): target values for energies.
-        p_e (np.ndarray): prediction values for forces.
-        y_f (np.ndarray): target values for energies.
-        p_f (np.ndarray): target values for forces.
+        p_e (np.ndarray): prediction values for energies.
+        y_f (np.ndarray): target values for forces.
+        p_f (np.ndarray): prediction values for forces.
     """
     if table_names is None:
         _, _, table_names, _ = io.analyze_hdf_tables(filename)

diff --git a/uf3/representation/bspline.py b/uf3/representation/bspline.py
@@ -355,9 +355,9 @@ def get_regularization_matrix(self,
 
         Args:
             ridge_map (dict): n-body term ridge regularizer strengths.
-                default: {1: 1e-4, 2: 1e-6, 3: 1e-5}
+                default: {1: 1e-8, 2: 0e0, 3: 1e-5}
             curvature_map (dict): n-body term curvature regularizer strengths.
-                default: {1: 0.0, 2: 1e-5, 3: 1e-5}
+                default: {1: 0.0, 2: 1e-8, 3: 1e-8}
 
         TODO: refactor to break up into smaller, reusable functions
 

diff --git a/uf3/representation/process.py b/uf3/representation/process.py
@@ -308,7 +308,7 @@ def evaluate_configuration(self,
                 {(name, 'e'), (name, 'fx'), ...{ instead of {'e', 'fx', ...}
             energy (float): energy of configuration (optional).
             forces (list, np.ndarray): array containing force components
-                fx, fy, fz for each atom. Expected shape is (n_atoms, 3).
+                fx, fy, fz for each atom. Expected shape is (3, n_atoms).
             energy_key (str): column name for energies, default "energy".
 
         Returns:
@@ -527,6 +527,7 @@ def get_training_tuples(self, df_features, kappa, data_coordinator):
             y (np.ndarray): target vector.
             w (np.ndarray): weight vector for machine learning.
         """
+        warnings.warn("get_training_tuples() is deprecated.", DeprecationWarning)
         energy_key = data_coordinator.energy_key
         x, y, w = dataframe_to_training_tuples(df_features,
                                                kappa=kappa,
@@ -626,4 +627,4 @@ def flatten_by_interactions(vector_map, pair_tuples):
         (np.ndarray): concatenated result of joining vector_map
             in order of occurrence in pair_tuples.
     """
-    return np.concatenate([vector_map[pair] for pair in pair_tuples], axis=-1)
+    return np.concatenate([vector_map[pair] for pair in pair_tuples], axis=-1)