Update files from the datasets library (from 1.7.0)

Release notes: https://github.com/huggingface/datasets/releases/tag/1.7.0

README.md

@@ -22,6 +22,7 @@ task_ids:
 - multi-class-classification
 - multi-label-classification
 - text-classification-other-emotion
+paperswithcode_id: goemotions
 ---
 
 # Dataset Card for GoEmotions
@@ -29,12 +30,12 @@ task_ids:
 ## Table of Contents
 - [Dataset Description](#dataset-description)
   - [Dataset Summary](#dataset-summary)
-  - [Supported Tasks](#supported-tasks-and-leaderboards)
+  - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
   - [Languages](#languages)
 - [Dataset Structure](#dataset-structure)
   - [Data Instances](#data-instances)
-  - [Data Fields](#data-instances)
-  - [Data Splits](#data-instances)
+  - [Data Fields](#data-fields)
+  - [Data Splits](#data-splits)
 - [Dataset Creation](#dataset-creation)
   - [Curation Rationale](#curation-rationale)
   - [Source Data](#source-data)

go_emotions.py

@@ -142,17 +142,17 @@ class GoEmotions(datasets.GeneratorBasedBuilder):
 
     def _generate_examples(self, filepaths, raw=False):
         """Generate AG News examples."""
-        for filepath in filepaths:
+        for file_idx, filepath in enumerate(filepaths):
             with open(filepath, "r", encoding="utf-8") as f:
                 if raw:
                     reader = csv.DictReader(f)
                 else:
                     reader = csv.DictReader(f, delimiter="\t", fieldnames=list(self.config.features.keys()))
 
-                for irow, row in enumerate(reader):
+                for row_idx, row in enumerate(reader):
                     if raw:
                         row["example_very_unclear"] = row["example_very_unclear"] == "TRUE"
                     else:
                         row["labels"] = [int(ind) for ind in row["labels"].split(",")]
 
-                    yield irow, row
+                    yield f"{file_idx}_{row_idx}", row