replaced 2**n with 1<<n

src/python/twitter/deepbird/projects/timelines/configs/rectweet_earlybird/feature_config.py

@@ -47,14 +47,8 @@ def get_feature_config(data_spec_path: str, label: str) -> FeatureConfigBuilder:
                 ("extended_encoded_tweet_features.label_spam_hi_rcl_flag", "A"),
                 ("extended_encoded_tweet_features.periscope_exists", "A"),
                 ("extended_encoded_tweet_features.periscope_has_been_featured", "A"),
-                (
-                    "extended_encoded_tweet_features.periscope_is_currently_featured",
-                    "A",
-                ),
-                (
-                    "extended_encoded_tweet_features.periscope_is_from_quality_source",
-                    "A",
-                ),
+                ("extended_encoded_tweet_features.periscope_is_currently_featured","A"),
+                ("extended_encoded_tweet_features.periscope_is_from_quality_source", "A"),
                 ("extended_encoded_tweet_features.periscope_is_live", "A"),
                 ("extended_encoded_tweet_features.quote_count", "A"),
                 ("extended_encoded_tweet_features.reply_count_v2", "A"),

src/python/twitter/deepbird/projects/timelines/scripts/models/earlybird/lolly/score.py

@@ -6,9 +6,6 @@ from .scorer import LollyModelScorer
 
 if __name__ == "__main__":
     lolly_model_reader = LollyModelReader(lolly_model_file_path=sys.argv[1])
-    lolly_model_scorer = LollyModelScorer(
-        data_example_parser=DBv2DataExampleParser(lolly_model_reader)
-    )
-
+    lolly_model_scorer = LollyModelScorer(DBv2DataExampleParser(lolly_model_reader))
     score = lolly_model_scorer.score(data_example=sys.argv[2])
     print(score)

src/python/twitter/deepbird/projects/timelines/scripts/models/earlybird/tf_model/discretizer_builder.py

@@ -40,9 +40,7 @@ class TFModelDiscretizerBuilder(object):
         :return: a HashingDiscretizer instance.
         """
         discretized_features = tf_model_initializer["features"]["discretized"]
-
         max_bins = 0
-
         feature_ids = []
         bin_vals = []
         for feature_name in discretized_features:
@@ -53,12 +51,9 @@ class TFModelDiscretizerBuilder(object):
                 np.float(bin_boundary) for bin_boundary in bin_boundaries
             ]
             bin_vals.append(np_bin_boundaries)
-
             max_bins = max(max_bins, len(np_bin_boundaries))
-
         feature_ids_np = np.array(feature_ids)
         bin_vals_np = np.array(bin_vals).flatten()
-
         return HashingDiscretizer(
             feature_ids=feature_ids_np,
             bin_vals=bin_vals_np,

src/python/twitter/deepbird/projects/timelines/scripts/models/earlybird/tf_model/hashing_utils.py

@@ -30,4 +30,4 @@ def make_feature_id(name: str, num_bits: int) -> np.int64:
 def limit_bits(value: int, num_bits: int) -> int:
     """Limits the number of bits in the given value."""
 
-    return value & ((2**num_bits) - 1)
+    return value & ((1<<num_bits) - 1)

src/python/twitter/deepbird/projects/timelines/scripts/models/earlybird/tf_model/weights_initializer_builder.py

@@ -16,9 +16,10 @@ class TFModelWeightsInitializerBuilder(object):
         """
         :return: (bias_initializer, weight_initializer)
         """
-        initial_weights = np.zeros((2**self.num_bits, 1))
 
+        initial_weights = np.zeros((1 << self.num_bits, 1))
         features = tf_model_initializer["features"]
+
         self._set_binary_feature_weights(initial_weights, features["binary"])
         self._set_discretized_feature_weights(initial_weights, features["discretized"])
 
@@ -27,8 +28,12 @@ class TFModelWeightsInitializerBuilder(object):
         ), twml.contrib.initializers.PartitionConstant(initial_weights)
 
     def _set_binary_feature_weights(
-        self, initial_weights: np.ndarray, binary_features: Dict[str, float]
+        self,
+        initial_weights: np.ndarray,
+        binary_features: Dict[str, float],
     ) -> None:
+        """set weights for binary features"""
+
         for feature_name, weight in binary_features.items():
             feature_id = make_feature_id(feature_name, self.num_bits)
             initial_weights[feature_id][0] = weight
@@ -38,6 +43,8 @@ class TFModelWeightsInitializerBuilder(object):
         initial_weights: np.ndarray,
         discretized_features: Dict[str, Dict[str, Any]],
     ) -> None:
+        """set weights for discretized features"""
+
         for feature_name, discretized_feature in discretized_features.items():
             feature_id = make_feature_id(feature_name, self.num_bits)
             for bin_idx, weight in enumerate(discretized_feature["weights"]):

twml/twml/contrib/layers/factorization_machine.py

@@ -89,7 +89,7 @@ class FactorizationMachine(Layer):
         # 2**30 is 1GB. 2 * (2**30) is 2GB.
         dtype = tf.as_dtype(self.dtype)
         requested_size = input_shape[1] * self.num_latent_variables * dtype.size
-        if requested_size >= 2**31:
+        if requested_size >= (1<<31):
             raise ValueError(
                 "Weight tensor can not be larger than 2GB. "
                 % "Requested Dimensions(%d, %d) of type %s (%d bytes total)"(

twml/twml/layers/full_sparse.py

@@ -187,7 +187,7 @@ class FullSparse(Layer):
         requested_size = (
             math.ceil(float(split_dim) / num_partitions) * other_dim * dtype.size
         )
-        if requested_size >= 2**31:
+        if requested_size >= 1 << 31:
             raise ValueError(
                 "Weight tensor partitions cannot be larger than 2GB.\n"
                 "Requested Dimensions(%d, %d) of type %s (%d bytes total) over %d partitions.\n"