add GA example in run.py (#24)

benchmarks/matbench_v0.1_modnet_v0.1.10/run.py

@@ -8,6 +8,8 @@
 from modnet.models import EnsembleMODNetModel
 from pymatgen.core import Composition
 
+USE_GA = False # wheter to use the GA or fit_preset (dynamic grid-search) for hyper-paremeter optimization.
+
 mb = MatbenchBenchmark(
     autoload=False, 
     subset=[
@@ -19,7 +21,11 @@
         'matbench_log_gvrh',
         'matbench_log_kvrh',
         'matbench_glass', 
-        'matbench_expt_is_metal', 
+        'matbench_expt_is_metal',
+        #'matbench_perovskites', # for the bigger tasks, USE_GA=True is recommended, as training time scales better with larger training sets
+        #'matbench_mp_e_form',
+        #'matbench_mp_gap',
+        #'matbench_mp_is_metal',
     ],
 )
 
@@ -65,18 +71,32 @@
         model = EnsembleMODNetModel(targets_hierarchy, weights)
 
         # fit model
-        (
-            models,
-            val_losses,
-            best_learning_curve,
-            learning_curves,
-            best_presets,
-        ) = model.fit_preset(
-            train_data,
-            classification=classification,
-            nested=5,
-            n_jobs=16,
-        )
+
+        if USE_GA:
+            # you can either use a GA for hyper-parameter optimization or...
+            from modnet.hyper_opt import FitGenetic
+            ga = FitGenetic(train_data)
+            model = ga.run(
+                size_pop=20,
+                num_generations=10,
+                n_jobs=16,
+                early_stopping=True,
+                refit=True,
+            )
+        else:
+            # ... a list of presets (kind of dynamic grid search)
+            (
+                models,
+                val_losses,
+                best_learning_curve,
+                learning_curves,
+                best_presets,
+            ) = model.fit_preset(
+                train_data,
+                classification=classification,
+                nested=5,
+                n_jobs=16,
+            )
 
         # Load and featurize test dataset
         test_df = task.get_test_data(fold, include_target=False, as_type="df")