Rename "Trasformer" to "Transformer"; Python 3 syntax updates (#1146)

* Rename "Trasformer" to "Transformer"

* Remove Python-2 style 'super' syntax

* Remove unneeded extension of object

* Update optimization_modules.py

ludwig/cli.py

@@ -22,7 +22,7 @@
 ludwig.contrib.contrib_import()
 
 
-class CLI(object):
+class CLI:
     """CLI describes a command line interface for interacting with Ludwig, there
     are several different functions that can be performed. These functions are:
     - experiment - run an experiment using ludwig

ludwig/datasets/amazon_review_polarity/__init__.py

@@ -40,7 +40,7 @@ def __init__(self, cache_dir=DEFAULT_CACHE_LOCATION):
         super().__init__(dataset_name="amazon_review_polarity", cache_dir=cache_dir)
 
     def process_downloaded_dataset(self):
-        super(AmazonPolarity, self).process_downloaded_dataset(header=None)
+        super().process_downloaded_dataset(header=None)
         processed_df = pd.read_csv(os.path.join(self.processed_dataset_path,
                                                 self.csv_filename))
         processed_df.columns = ['label', 'review_tile', 'review_text', 'split']

ludwig/datasets/amazon_reviews/__init__.py

@@ -40,7 +40,7 @@ def __init__(self, cache_dir=DEFAULT_CACHE_LOCATION):
         super().__init__(dataset_name="amazon_reviews", cache_dir=cache_dir)
 
     def process_downloaded_dataset(self):
-        super(AmazonReviews, self).process_downloaded_dataset(header=None)
+        super().process_downloaded_dataset(header=None)
         processed_df = pd.read_csv(os.path.join(self.processed_dataset_path,
                                                 self.csv_filename))
         processed_df.columns = ['label', 'review_tile', 'review_text', 'split']

ludwig/datasets/dbpedia/__init__.py

@@ -38,7 +38,7 @@ def __init__(self, cache_dir=DEFAULT_CACHE_LOCATION):
         super().__init__(dataset_name="dbpedia", cache_dir=cache_dir)
 
     def process_downloaded_dataset(self):
-        super(DBPedia, self).process_downloaded_dataset(header=None)
+        super().process_downloaded_dataset(header=None)
         processed_df = pd.read_csv(os.path.join(self.processed_dataset_path,
                                                 self.csv_filename))
         processed_df.columns = ['label', 'title', 'content', 'split']

ludwig/datasets/ethos_binary/__init__.py

@@ -36,7 +36,7 @@ def __init__(self, cache_dir=DEFAULT_CACHE_LOCATION):
         super().__init__(dataset_name="ethos_binary", cache_dir=cache_dir)
 
     def process_downloaded_dataset(self):
-        super(EthosBinary, self).process_downloaded_dataset()
+        super().process_downloaded_dataset()
         # replace ; sperator to ,
         processed_df = pd.read_csv(os.path.join(self.processed_dataset_path,
                                                 self.csv_filename), sep=";")

ludwig/datasets/goemotions/__init__.py

@@ -43,7 +43,7 @@ def read_file(self, filetype, filename):
         return file_df
 
     def process_downloaded_dataset(self):
-        super(GoEmotions, self).process_downloaded_dataset()
+        super().process_downloaded_dataset()
         # format emotion ids to be a set of emotion ids vs. string
         processed_df = pd.read_csv(os.path.join(self.processed_dataset_path,
                                                 self.csv_filename))

ludwig/datasets/sst2/sst_utils.py

@@ -173,7 +173,7 @@ def process_downloaded_dataset(self):
                 index=False
             )
 
-        super(SST, self).process_downloaded_dataset()
+        super().process_downloaded_dataset()
 
 
 def format_text(text: str):

ludwig/datasets/yahoo_answers/__init__.py

@@ -40,7 +40,7 @@ def __init__(self, cache_dir=DEFAULT_CACHE_LOCATION):
         super().__init__(dataset_name="yahoo_answers", cache_dir=cache_dir)
 
     def process_downloaded_dataset(self):
-        super(YahooAnswers, self).process_downloaded_dataset(header=None)
+        super().process_downloaded_dataset(header=None)
         processed_df = pd.read_csv(os.path.join(self.processed_dataset_path,
                                                 self.csv_filename))
         processed_df.columns = ['label', 'question_title', 'question', 'best_answer', 'split']

ludwig/datasets/yelp_review_polarity/__init__.py

@@ -39,7 +39,7 @@ def __init__(self, cache_dir=DEFAULT_CACHE_LOCATION):
         super().__init__(dataset_name="yelp_polarity", cache_dir=cache_dir)
 
     def process_downloaded_dataset(self):
-        super(YelpPolarity, self).process_downloaded_dataset(header=None)
+        super().process_downloaded_dataset(header=None)
         processed_df = pd.read_csv(os.path.join(self.processed_dataset_path,
                                                 self.csv_filename))
         processed_df.columns = ['label', 'text', 'split']

ludwig/datasets/yelp_reviews/__init__.py

@@ -39,7 +39,7 @@ def __init__(self, cache_dir=DEFAULT_CACHE_LOCATION):
         super().__init__(dataset_name="yelp_reviews", cache_dir=cache_dir)
 
     def process_downloaded_dataset(self):
-        super(YelpReviews, self).process_downloaded_dataset(header=None)
+        super().process_downloaded_dataset(header=None)
         processed_df = pd.read_csv(os.path.join(self.processed_dataset_path,
                                                 self.csv_filename))
         processed_df.columns = ['label', 'text', 'split']

ludwig/decoders/sequence_decoders.py

@@ -78,7 +78,7 @@ def __init__(
             reduce_input='sum',
             **kwargs
     ):
-        super(SequenceGeneratorDecoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         self.cell_type = cell_type
@@ -720,7 +720,7 @@ def __init__(
             is_timeseries=False,
             **kwargs
     ):
-        super(SequenceTaggerDecoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         self.attention = attention

ludwig/encoders/bag_encoders.py

@@ -62,7 +62,7 @@ def __init__(
             dropout=0.0,
             **kwargs
     ):
-        super(BagEmbedWeightedEncoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         logger.debug('  EmbedWeighted')

ludwig/encoders/binary_encoders.py

@@ -45,7 +45,7 @@ def __init__(
             self,
             **kwargs
     ):
-        super(BinaryPassthroughEncoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
     def call(self, inputs, training=None, mask=None):

ludwig/encoders/category_encoders.py

@@ -51,7 +51,7 @@ def __init__(
             embedding_regularizer=None,
             **kwargs
     ):
-        super(CategoricalEmbedEncoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         logger.debug('  Embed')
@@ -96,7 +96,7 @@ def __init__(
             embedding_regularizer=None,
             **kwargs
     ):
-        super(CategoricalSparseEncoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         logger.debug('  Embed')

ludwig/encoders/date_encoders.py

@@ -109,7 +109,7 @@ def __init__(
             :type dropout: float
 
         """
-        super(DateEmbed, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         logger.debug('  year FCStack')
@@ -379,7 +379,7 @@ def __init__(
                    returning the encoder output.
             :type dropout: float
         """
-        super(DateWave, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         logger.debug('  year FCStack')

ludwig/encoders/generic_encoders.py

@@ -29,7 +29,7 @@ def __init__(
             self,
             **kwargs
     ):
-        super(PassthroughEncoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
     def call(self, inputs, training=None, mask=None):
@@ -61,7 +61,7 @@ def __init__(
             dropout=0,
             **kwargs
     ):
-        super(DenseEncoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         logger.debug('  FCStack')

ludwig/encoders/h3_encoders.py

@@ -102,7 +102,7 @@ def __init__(
                    is greater than 0).
             :type regularize: Boolean
         """
-        super(H3Embed, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         self.embedding_size = embedding_size
@@ -333,7 +333,7 @@ def __init__(
                    is greater than 0).
             :type regularize: Boolean
         """
-        super(H3WeightedSum, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         self.should_softmax = should_softmax
@@ -540,7 +540,7 @@ def __init__(
                    (which does not reduce and returns the full tensor).
             :type reduce_output: str
         """
-        super(H3RNN, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         self.embedding_size = embedding_size

ludwig/encoders/image_encoders.py

@@ -83,7 +83,7 @@ def __init__(
             fc_dropout=0,
             **kwargs
     ):
-        super(Stacked2DCNN, self).__init__()
+        super().__init__()
 
         logger.debug(' {}'.format(self.name))
 
@@ -183,7 +183,7 @@ def __init__(
             dropout=0,
             **kwargs
     ):
-        super(ResNetEncoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         if resnet_size < 50:

ludwig/encoders/sequence_encoders.py

@@ -22,7 +22,7 @@
 
 from ludwig.encoders.base import Encoder
 from ludwig.utils.registry import Registry, register, register_default
-from ludwig.modules.attention_modules import TrasformerStack
+from ludwig.modules.attention_modules import TransformerStack
 from ludwig.modules.convolutional_modules import Conv1DStack, \
     ParallelConv1DStack, ParallelConv1D
 from ludwig.modules.embedding_modules import EmbedSequence, \
@@ -61,7 +61,7 @@ def __init__(
                    and returns the full tensor).
             :type reduce_output: str
         """
-        super(SequencePassthroughEncoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         self.reduce_output = reduce_output
@@ -189,7 +189,7 @@ def __init__(
             :type dropout: Tensor
 
         """
-        super(SequenceEmbedEncoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         self.reduce_output = reduce_output
@@ -401,7 +401,7 @@ def __init__(
                    (which does not reduce and returns the full tensor).
             :type reduce_output: str
         """
-        super(ParallelCNN, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         if conv_layers is not None and num_conv_layers is None:
@@ -723,7 +723,7 @@ def __init__(
                    (which does not reduce and returns the full tensor).
             :type reduce_output: str
         """
-        super(StackedCNN, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         if conv_layers is not None and num_conv_layers is None:
@@ -1083,7 +1083,7 @@ def __init__(
                    (which does not reduce and returns the full tensor).
             :type reduce_output: str
         """
-        super(StackedParallelCNN, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         if stacked_layers is not None and num_stacked_layers is None:
@@ -1412,7 +1412,7 @@ def __init__(
                    (which does not reduce and returns the full tensor).
             :type reduce_output: str
         """
-        super(StackedRNN, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         self.reduce_output = reduce_output
@@ -1676,7 +1676,7 @@ def __init__(
                    (which does not reduce and returns the full tensor).
             :type reduce_output: str
         """
-        super(StackedCNNRNN, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         if conv_layers is not None and num_conv_layers is None:
@@ -2005,7 +2005,7 @@ def __init__(
                    (which does not reduce and returns the full tensor).
             :type reduce_output: str
         """
-        super(StackedTransformer, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         self.reduce_output = reduce_output
@@ -2042,7 +2042,7 @@ def __init__(
             self.should_project = True
 
         logger.debug('  TransformerStack')
-        self.transformer_stack = TrasformerStack(
+        self.transformer_stack = TransformerStack(
             hidden_size=hidden_size,
             num_heads=num_heads,
             fc_size=transformer_fc_size,

ludwig/encoders/set_encoders.py

@@ -63,7 +63,7 @@ def __init__(
             reduce_output='sum',
             **kwargs
     ):
-        super(SetSparseEncoder, self).__init__()
+        super().__init__()
         logger.debug(' {}'.format(self.name))
 
         logger.debug('  EmbedSparse')