suggestions from review

Author: ravinkohli

autoPyTorch/pipeline/base_pipeline.py

@@ -300,8 +300,7 @@ def _get_hyperparameter_search_space(self,
     def _add_forbidden_conditions(self, cs: ConfigurationSpace) -> ConfigurationSpace:
         """
         Add forbidden conditions to ensure valid configurations.
-        Currently, Learned Entity Embedding is only valid when encoder is one hot encoder
-        and CyclicLR is disabled when using stochastic weight averaging and snapshot
+        Currently, CyclicLR is disabled when using stochastic weight averaging and snapshot
         ensembling.
 
         Args:

autoPyTorch/pipeline/components/setup/network_embedding/LearnedEntityEmbedding.py

@@ -23,8 +23,7 @@ def __init__(self, config: Dict[str, Any], num_categories_per_col: np.ndarray, n
         """
         Args:
             config (Dict[str, Any]): The configuration sampled by the hyperparameter optimizer
-            num_input_features (np.ndarray): column wise information of number of output columns after transformation
-                for each categorical column and 0 for numerical columns
+            num_categories_per_col (np.ndarray): number of categories per categorical columns that will be embedded
             num_features_excl_embed (int): number of features in X excluding the features that need to be embedded
         """
         super().__init__()

autoPyTorch/pipeline/tabular_classification.py

@@ -51,20 +51,21 @@ class TabularClassificationPipeline(ClassifierMixin, BasePipeline):
     It implements a pipeline, which includes the following as steps:
 
     1. `imputer`
-    2. `encoder`
-    3. `scaler`
-    4. `feature_preprocessor`
-    5. `tabular_transformer`
-    6. `preprocessing`
-    7. `network_embedding`
-    8. `network_backbone`
-    9. `network_head`
-    10. `network`
-    11. `network_init`
-    12. `optimizer`
-    13. `lr_scheduler`
-    14. `data_loader`
-    15. `trainer`
+    2. `column_splitter
+    3. `encoder`
+    4. `scaler`
+    5. `feature_preprocessor`
+    6. `tabular_transformer`
+    7. `preprocessing`
+    8. `network_embedding`
+    9. `network_backbone`
+    10. `network_head`
+    11. `network`
+    12. `network_init`
+    13. `optimizer`
+    14. `lr_scheduler`
+    15. `data_loader`
+    16. `trainer`
 
     Contrary to the sklearn API it is not possible to enumerate the
     possible parameters in the __init__ function because we only know the

autoPyTorch/pipeline/tabular_regression.py

@@ -53,20 +53,21 @@ class TabularRegressionPipeline(RegressorMixin, BasePipeline):
     It implements a pipeline, which includes the following as steps:
 
     1. `imputer`
-    2. `encoder`
-    3. `scaler`
-    4. `feature_preprocessor`
-    5. `tabular_transformer`
-    6. `preprocessing`
-    7. `network_embedding`
-    8. `network_backbone`
-    9. `network_head`
-    10. `network`
-    11. `network_init`
-    12. `optimizer`
-    13. `lr_scheduler`
-    14. `data_loader`
-    15. `trainer`
+    2. `column_splitter
+    3. `encoder`
+    4. `scaler`
+    5. `feature_preprocessor`
+    6. `tabular_transformer`
+    7. `preprocessing`
+    8. `network_embedding`
+    9. `network_backbone`
+    10. `network_head`
+    11. `network`
+    12. `network_init`
+    13. `optimizer`
+    14. `lr_scheduler`
+    15. `data_loader`
+    16. `trainer`
 
     Contrary to the sklearn API it is not possible to enumerate the
     possible parameters in the __init__ function because we only know the