Command-line options for LAMBO segmentation.

combo/main.py

@@ -88,8 +88,8 @@ flags.DEFINE_boolean(name="silent", default=True,
 flags.DEFINE_enum(name="predictor_name", default="combo-spacy",
                   enum_values=["combo", "combo-spacy", "combo-lambo"],
                   help="Use predictor with whitespace, spacy or LAMBO tokenizer.")
-flags.DEFINE_string(name="lambo_model_name", default="en",
-                    help="LAMBO model name (if LAMBO used for segmentation).")
+#flags.DEFINE_string(name="lambo_model_name", default="en",
+#                    help="LAMBO model name (if LAMBO used for segmentation).")
 
 def run(_):
     """Run model."""
@@ -176,7 +176,7 @@ def _get_predictor() -> predictors.Predictor:
     )
 
     return predictors.Predictor.from_archive(
-        archive, FLAGS.predictor_name, "validation", True, {"lambo_model_name" : FLAGS.lambo_model_name}
+        archive, FLAGS.predictor_name#, extra_args={"lambo_model_name" : FLAGS.lambo_model_name}
     )
 
 

combo/predict.py

@@ -238,7 +238,7 @@ class COMBO(predictor.Predictor):
     
     @classmethod
     def with_lambo_tokenizer(cls, model: models.Model,
-                             dataset_reader: allen_data.DatasetReader, lambo_model_name : str):
+                             dataset_reader: allen_data.DatasetReader, lambo_model_name : str = 'en'):
         return cls(model, dataset_reader, lambo.LamboTokenizer(lambo_model_name))
 
     @classmethod