Merge pull request #25 from szepeviktor/typos

Fix typos

Author: dholzmueller

docs/source/bench/03_code.md

@@ -14,7 +14,7 @@ a list of the most important hyperparameters.
 
 ## Datasets
 
-We represent our datasets using the `DictDatset` class from `tab_models`.
+We represent our datasets using the `DictDataset` class from `tab_models`.
 These datasets can be loaded as follows:
 
 ```python

docs/source/models/01_sklearn_interfaces.rst

@@ -112,7 +112,7 @@ but fails to run predict() due to pytorch-lightning device issues.
 
 .. code-block:: language
     import torch
-    import dill  # might also work with pickle instad
+    import dill  # might also work with pickle instead
     torch.save(model, 'model.pkl', pickle_module=dill, _use_new_zipfile_serialization=False)
     model = torch.load('model.pkl', map_location='cpu', pickle_module=dill)
 

docs/source/models/nn_classes.md

@@ -35,7 +35,7 @@ for continuous and categorical variables.
 Moreover, during training, we also pass the labels 'y' through the Layer, 
 which allows to implement mixup, label smoothing, 
 and output standardization as Layer objects.
-- `Fitter` classes initialize the NN based on a single foward pass 
+- `Fitter` classes initialize the NN based on a single forward pass 
 on the (subsampled) training (and possibly validation) set. 
 This is done using the `fit()` or `fit_transform()` functions 
 similar to scikit-learn preprocessing classes, 

pytabkit/bench/data/get_uci.py

@@ -621,7 +621,7 @@ def get_metro_interstate_traffic_volume():
     replace_chars_in_file('metro_interstate_traffic_volume.data', '  ', ' ')
 
 
-    # Now we are in the position ot read the data, convert the time and date, and movel the labels
+    # Now we are in the position to read the data, convert the time and date, and movel the labels
 
 
     data = load_raw_data('metro_interstate_traffic_volume_short.data', ',', description_columns = 0, date_column = 16, date_sep = '-', date_order = 'Ymd', time_column = 17, time_sep = ':')
@@ -736,7 +736,7 @@ def get_tarvel_review_ratings():
     prepare_new_data_set_group_id()
 
 
-    # Download the data and correct the mispelling of its name
+    # Download the data and correct the misspelling of its name
 
     download_and_save('http://archive.ics.uci.edu/ml/machine-learning-databases/00485/google_review_ratings.csv', 'travel_review_ratings.data')
 
@@ -1332,8 +1332,8 @@ def get_smartphone_human_activity():
 def get_artificial_characters():
 
     prepare_new_data_set_group_id()
-    #download_and_save('https://archive.ics.uci.edu/ml/machine-learning-databases/artificial-characters/character.tar.Z', 'artficial_characters.tar.Z')
-    #download_and_save('https://archive.ics.uci.edu/ml/machine-learning-databases/artificial-characters/character.names', 'artficial_characters.description')
+    #download_and_save('https://archive.ics.uci.edu/ml/machine-learning-databases/artificial-characters/character.tar.Z', 'artificial_characters.tar.Z')
+    #download_and_save('https://archive.ics.uci.edu/ml/machine-learning-databases/artificial-characters/character.names', 'artificial_characters.description')
 
     print("Artificial Characters is currently not processed since:")
     print("  - the data comes in a rather convoluted form")
@@ -1534,7 +1534,7 @@ def get_thyroids():
     download_and_save('https://archive.ics.uci.edu/ml/machine-learning-databases/thyroid-disease/dis.test', 'thyroid_dis.test.data') 
     download_and_save('https://archive.ics.uci.edu/ml/machine-learning-databases/thyroid-disease/dis.names', 'thyroid_dis.description')
 
-    # new-thyroid.data only contains 215 samples and is thus ommitted
+    # new-thyroid.data only contains 215 samples and is thus omitted
 
     download_and_save('http://archive.ics.uci.edu/ml/machine-learning-databases/thyroid-disease/hypothyroid.data', 'thyroid_hypo.data') 
     download_and_save('http://archive.ics.uci.edu/ml/machine-learning-databases/thyroid-disease/hypothyroid.names', 'thyroid_hypo.description')
@@ -2819,7 +2819,7 @@ def get_wave_energy():
 
     # For each of the 4 data sets, the last column contains the sum of columns 32 to 47.
     # I assume the last column is the label and columns 32 to 47 are intermediate results
-    # and that only the first 32 colums should be used as features.
+    # and that only the first 32 columns should be used as features.
     indices = range(32, 48)
 
     data_adelaide = load_raw_data('WECs_DataSet/Adelaide_Data.csv', sep=',')

pytabkit/bench/scheduling/jobs.py

@@ -17,12 +17,12 @@ def __init__(self, job_id: int, time_s: float,
         """
         :param job_id: Job id.
         :param time_s: Time in seconds that the job ran for.
-        :param oom_cpu: Whether an out-of-memory error occured on the CPU.
-        :param oom_gpu: Whether an out-of-memory error occured on the GPU.
+        :param oom_cpu: Whether an out-of-memory error occurred on the CPU.
+        :param oom_gpu: Whether an out-of-memory error occurred on the GPU.
         :param finished_normally: Whether the job ran normally,
             such that its time and RAM values are representative of how it would normally run.
             For example, if the job ran faster because the results were already partially precomputed,
-            it should not count towards the time estimation. Of course, if an exception occured,
+            it should not count towards the time estimation. Of course, if an exception occurred,
             we should have finished_normally=False.
         :param exception_msg: Exception message (if there was any).
         """

pytabkit/models/alg_interfaces/catboost_interfaces.py

@@ -252,7 +252,7 @@ def _fit(self, train_ds: DictDataset, val_ds: Optional[DictDataset], params: Dic
             # with these parameters, catboost will reload from the model automatically if it is there
         bst = catboost.CatBoost(params)
         with warnings.catch_warnings():
-            warnings.filterwarnings('ignore', message='Can\'t optimze method "evaluate" because self argument is used')
+            warnings.filterwarnings('ignore', message='Can\'t optimize method "evaluate" because self argument is used')
             bst.fit(self._convert_ds(train_ds), eval_set=None if val_ds is None else self._convert_ds(val_ds))
 
         if val_ds is not None:

pytabkit/models/alg_interfaces/nn_interfaces.py

@@ -307,7 +307,7 @@ def __init__(self, space: Optional[Union[str, Dict[str, Any]]] = None, n_hyperop
             }
             utils.update_dict(default_config, remove_keys=list(space.keys()))
         elif not isinstance(space, dict):
-            print(f'Unkown hyperparameter space: {space}')
+            print(f'Unknown hyperparameter space: {space}')
 
         config = utils.update_dict(default_config, config)
         opt_class = SMACOptimizer if opt_method == 'smac' else HyperoptOptimizer

pytabkit/models/alg_interfaces/other_interfaces.py

@@ -1053,7 +1053,7 @@ def fit(self, X, y, X_val, y_val, cat_features: Optional[List[str]] = None):
             'learning_rate_weights': 0.005,  # learning rate for leaf weights
             'learning_rate_index': 0.01,  # learning rate for split indices
             'learning_rate_values': 0.01,  # learning rate for split values
-            'learning_rate_leaf': 0.01,  # learning rate for leafs (logits)
+            'learning_rate_leaf': 0.01,  # learning rate for leaves (logits)
 
             'optimizer': 'adam',  # optimizer
             'cosine_decay_steps': 0,  # decay steps for lr schedule (CosineDecayRestarts)

pytabkit/models/alg_interfaces/tabm_interface.py

@@ -334,7 +334,7 @@ def evaluate(part: str) -> float:
         }
         best_params = [p.clone() for p in model.parameters()]
         # Early stopping: the training stops when
-        # there are more than `patience` consequtive bad updates.
+        # there are more than `patience` consecutive bad updates.
         remaining_patience = patience
 
         try:

pytabkit/models/alg_interfaces/xgboost_interfaces.py

@@ -498,7 +498,7 @@ def _sample_params(self, is_classification: bool, seed: int, n_train: int):
                 # 'max_bin'  # also makes things slower
             }
         elif hpo_space_name == 'large-v3':
-            # shrunk verion of large-v2: removed gamma, colsample_bytree
+            # shrunk version of large-v2: removed gamma, colsample_bytree
             params = {
                 'n_estimators': 1000,
                 'early_stopping_rounds': 50,
@@ -513,7 +513,7 @@ def _sample_params(self, is_classification: bool, seed: int, n_train: int):
                 'grow_policy': rng.choice(['depthwise', 'lossguide']),
             }
         elif hpo_space_name == 'large-v4':
-            # modified verion of large-v3
+            # modified version of large-v3
             params = {
                 'n_estimators': 1000,
                 'early_stopping_rounds': 50,
@@ -529,7 +529,7 @@ def _sample_params(self, is_classification: bool, seed: int, n_train: int):
                 'max_leaves': round(np.exp(rng.uniform(np.log(2.0), np.log(2048.0))))  # added
             }
         elif hpo_space_name == 'large-v5':
-            # shrunk verion of large-v3 but without the extra stuff from large-v4
+            # shrunk version of large-v3 but without the extra stuff from large-v4
             params = {
                 'n_estimators': 1000,
                 'early_stopping_rounds': 50,
@@ -543,7 +543,7 @@ def _sample_params(self, is_classification: bool, seed: int, n_train: int):
                 'reg_lambda': np.exp(rng.uniform(np.log(1e-3), np.log(20.0))),  # modified
             }
         elif hpo_space_name == 'large-v6':
-            # shrunk verion of large-v4
+            # shrunk version of large-v4
             params = {
                 'n_estimators': 1000,
                 'early_stopping_rounds': 50,