Adapt to AR_predictions changes

deepsphere · Aug 5, 2021 · b85757a · b85757a
1 parent ae55ff7
commit b85757a
Show file tree

Hide file tree

Showing 5 changed files with 14 additions and 9 deletions.
diff --git a/train_state.py b/train_state.py
@@ -475,7 +475,8 @@ def main(cfg_path, exp_dir, data_dir):
                                              da_dynamic = da_test_dynamic,
                                              da_static = da_static,              
                                              da_bc = da_test_bc, 
-                                             scaler = scaler,
+                                             scaler_transform = scaler,
+                                             scaler_inverse = scaler,
                                              # Dataloader options
                                              device = device,
                                              batch_size = 50,  # number of forecasts per batch

diff --git a/train_state_ensemble.py b/train_state_ensemble.py
@@ -484,7 +484,8 @@ def main(cfg_path, exp_dir, data_dir, nb_models):
                                                 da_dynamic = da_test_dynamic,
                                                 da_static = da_static,              
                                                 da_bc = da_test_bc, 
-                                                scaler = scaler,
+                                                scaler_transform = scaler,
+                                                scaler_inverse = scaler,
                                                 # Dataloader options
                                                 device = device,
                                                 batch_size = 50,  # number of forecasts per batch

diff --git a/train_state_increment.py b/train_state_increment.py
@@ -475,7 +475,8 @@ def main(cfg_path, exp_dir, data_dir):
                                              da_dynamic = da_test_dynamic,
                                              da_static = da_static,              
                                              da_bc = da_test_bc, 
-                                             scaler = scaler,
+                                             scaler_transform = scaler,
+                                             scaler_inverse = scaler,
                                              # Dataloader options
                                              device = device,
                                              batch_size = 50,  # number of forecasts per batch

diff --git a/train_state_short.py b/train_state_short.py
@@ -475,7 +475,8 @@ def main(cfg_path, exp_dir, data_dir):
                                              da_dynamic = da_test_dynamic,
                                              da_static = da_static,              
                                              da_bc = da_test_bc, 
-                                             scaler = scaler,
+                                             scaler_transform = scaler,
+                                             scaler_inverse = scaler,
                                              # Dataloader options
                                              device = device,
                                              batch_size = 50,  # number of forecasts per batch

diff --git a/train_zarr_100km.py b/train_zarr_100km.py
@@ -220,13 +220,13 @@ def main(cfg_path, exp_dir, data_dir):
     print_model_description(cfg)
     print_dim_info(dim_info)  
 
-    from modules.utils_config import get_pytorch_model
-    model = get_pytorch_model(module_with_custom_models = my_architectures,
-                              model_settings = model_settings,
-                              training_settings = training_settings)
     ##------------------------------------------------------------------------.
     ### Define the model architecture  
     # - TODO: improve with utils_config.get_pytorch_model
+    # from modules.utils_config import get_pytorch_model
+    # model = get_pytorch_model(module_with_custom_models = my_architectures,
+    #                           model_settings = model_settings,
+    #                           training_settings = training_settings)
     DeepSphereModelClass = getattr(my_architectures, model_settings['architecture_name'])
     # - Retrieve required model arguments
     model_keys = ['dim_info', 'sampling', 'resolution',
@@ -449,7 +449,8 @@ def main(cfg_path, exp_dir, data_dir):
                                              da_dynamic = da_test_dynamic,
                                              da_static = da_static,              
                                              da_bc = da_test_bc, 
-                                             scaler = scaler,
+                                             scaler_transform = scaler,
+                                             scaler_inverse = scaler,
                                              # Dataloader options
                                              device = device,
                                              batch_size = 50,  # number of forecasts per batch