Datasets:

nlp-thedeep
/

humsetbias

Tasks:

Text Classification

Token Classification

Sub-tasks:

multi-label-classification

Size:

Tags:

analytical-framework

License:

Dataset card Files Files and versions Community

nlp-thedeep commited on May 11, 2023

Commit

f09fb05

·

1 Parent(s): 0dab569

Update humsetbias.py

Files changed (1) hide show

humsetbias.py +3 -3

humsetbias.py CHANGED Viewed

@@ -219,11 +219,11 @@ class Humset(datasets.GeneratorBasedBuilder):
             for line in data:
                 row = json.loads(line)
                 if "train" in filepath:
-                    row = {k: v for k, v in row if k in HUMSETBIAS_GENERAL}
                 elif "gender" in filepath:
-                    row = {k: v for k, v in row if k in HUMSETBIAS_FEATURES_GENDER}
                 elif "country" in filepath:
-                    row = {k: v for k, v in row if k in HUMSETBIAS_FEATURES_COUNTRY}
                 #if self.config.name == "1.0.0":
                 yield idx, row
                 idx+=1

             for line in data:
                 row = json.loads(line)
                 if "train" in filepath:
+                    row = {k: v for k, v in row.items() if k in HUMSETBIAS_GENERAL}
                 elif "gender" in filepath:
+                    row = {k: v for k, v in row.items() if k in HUMSETBIAS_FEATURES_GENDER}
                 elif "country" in filepath:
+                    row = {k: v for k, v in row.items() if k in HUMSETBIAS_FEATURES_COUNTRY}
                 #if self.config.name == "1.0.0":
                 yield idx, row
                 idx+=1