antypasd commited on
Commit
a9c8c25
1 Parent(s): 4e3e08b

added hate labels

Browse files
.gitignore CHANGED
@@ -1 +1,2 @@
1
  misc
 
 
1
  misc
2
+ dimos.ipynb
NOTE.md CHANGED
@@ -4,4 +4,4 @@
4
  - TweetNER: [link](https://huggingface.co/datasets/tner/tweetner7)
5
  - TweetQA: [link](https://huggingface.co/datasets/tweet_qa), [link2](https://huggingface.co/datasets/lmqg/qg_tweetqa)
6
  - TweetIntimacy: [link]()
7
-
 
4
  - TweetNER: [link](https://huggingface.co/datasets/tner/tweetner7)
5
  - TweetQA: [link](https://huggingface.co/datasets/tweet_qa), [link2](https://huggingface.co/datasets/lmqg/qg_tweetqa)
6
  - TweetIntimacy: [link]()
7
+ - TweetSentiment: [link](https://alt.qcri.org/semeval2017/task4/index.php?id=results)
data/tweet_nerd/test.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
data/tweet_nerd/train.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
get_stats.py CHANGED
@@ -9,6 +9,10 @@ task_description = {
9
  'tweet_similarity': "regression on two texts",
10
  'tweet_topic': "multi-label classification",
11
  "tempo_wic": "binary classification on two texts"
 
 
 
 
12
  }
13
  for task in task_description.keys():
14
  data = load_dataset("cardiffnlp/super_tweet_eval", task)
 
9
  'tweet_similarity': "regression on two texts",
10
  'tweet_topic': "multi-label classification",
11
  "tempo_wic": "binary classification on two texts"
12
+ "tweet_sentiment": "ABSA on a five-point scale"
13
+ "tweet_hate": "multi-class classification"
14
+ "tweet_emoji": "multi-class classification"
15
+ "tweet_nerd": "binary classification"
16
  }
17
  for task in task_description.keys():
18
  data = load_dataset("cardiffnlp/super_tweet_eval", task)
process/tweet_sentiment.py CHANGED
@@ -1,4 +1,3 @@
1
- # Original data: https://alt.qcri.org/semeval2017/task4/index.php?id=results
2
  import pandas as pd
3
  from glob import glob
4
  import urllib
 
 
1
  import pandas as pd
2
  from glob import glob
3
  import urllib
super_tweet_eval.py CHANGED
@@ -2,7 +2,7 @@
2
  import json
3
  import datasets
4
 
5
- _VERSION = "0.1.32"
6
  _SUPER_TWEET_EVAL_CITATION = """TBA"""
7
  _SUPER_TWEET_EVAL_DESCRIPTION = """TBA"""
8
  _TWEET_TOPIC_DESCRIPTION = """
@@ -279,11 +279,11 @@ class SuperTweetEval(datasets.GeneratorBasedBuilder):
279
  features["text_1_tokenized"] = datasets.Sequence(datasets.Value("string"))
280
  features["text_2_tokenized"] = datasets.Sequence(datasets.Value("string"))
281
  if self.config.name == "tweet_hate":
282
- names = [
283
- 'target_gender_aggregated','target_race_aggregated', 'target_sexuality_aggregated',
284
- 'target_religion_aggregated','target_origin_aggregated', 'target_disability_aggregated','target_age_aggregated',
285
- 'not_hate']
286
- features["gold_label"] = datasets.Value("int32")
287
  features["text"] = datasets.Value("string")
288
  if self.config.name == "tweet_nerd":
289
  features['target'] = datasets.Value("string")
@@ -308,11 +308,7 @@ class SuperTweetEval(datasets.GeneratorBasedBuilder):
308
  )
309
 
310
  def _split_generators(self, dl_manager):
311
- # TODO: temporary check until we have all splits online
312
- if self.config.name == 'tweet_nerd':
313
- splits = ['validation']
314
- else:
315
- splits = ['train', 'test', 'validation']
316
  downloaded_file = dl_manager.download_and_extract({s: f"{self.config.data_url}/{s}.jsonl" for s in splits})
317
  return [datasets.SplitGenerator(name=s, gen_kwargs={"filepath": downloaded_file[s]}) for s in splits]
318
 
 
2
  import json
3
  import datasets
4
 
5
+ _VERSION = "0.1.33"
6
  _SUPER_TWEET_EVAL_CITATION = """TBA"""
7
  _SUPER_TWEET_EVAL_DESCRIPTION = """TBA"""
8
  _TWEET_TOPIC_DESCRIPTION = """
 
279
  features["text_1_tokenized"] = datasets.Sequence(datasets.Value("string"))
280
  features["text_2_tokenized"] = datasets.Sequence(datasets.Value("string"))
281
  if self.config.name == "tweet_hate":
282
+ label_classes = [
283
+ 'hate_gender','hate_race', 'hate_sexuality', 'hate_religion','hate_origin', 'hate_disability',
284
+ 'target_age', 'not_hate']
285
+ features['gold_label'] = datasets.features.ClassLabel(names=self.config.label_classes)
286
+ #features["gold_label"] = datasets.Value("int32")
287
  features["text"] = datasets.Value("string")
288
  if self.config.name == "tweet_nerd":
289
  features['target'] = datasets.Value("string")
 
308
  )
309
 
310
  def _split_generators(self, dl_manager):
311
+ splits = ['train', 'test', 'validation']
 
 
 
 
312
  downloaded_file = dl_manager.download_and_extract({s: f"{self.config.data_url}/{s}.jsonl" for s in splits})
313
  return [datasets.SplitGenerator(name=s, gen_kwargs={"filepath": downloaded_file[s]}) for s in splits]
314