@thien commited on
Commit
1975910
1 Parent(s): cc4eb1a

initial commit

Browse files
.gitattributes CHANGED
@@ -53,3 +53,62 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
53
  *.jpg filter=lfs diff=lfs merge=lfs -text
54
  *.jpeg filter=lfs diff=lfs merge=lfs -text
55
  *.webp filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
53
  *.jpg filter=lfs diff=lfs merge=lfs -text
54
  *.jpeg filter=lfs diff=lfs merge=lfs -text
55
  *.webp filter=lfs diff=lfs merge=lfs -text
56
+ ||||||| (empty tree)
57
+ =======
58
+ *.7z filter=lfs diff=lfs merge=lfs -text
59
+ *.arrow filter=lfs diff=lfs merge=lfs -text
60
+ *.bin filter=lfs diff=lfs merge=lfs -text
61
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
62
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
63
+ *.ftz filter=lfs diff=lfs merge=lfs -text
64
+ *.gz filter=lfs diff=lfs merge=lfs -text
65
+ *.h5 filter=lfs diff=lfs merge=lfs -text
66
+ *.joblib filter=lfs diff=lfs merge=lfs -text
67
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
68
+ *.lz4 filter=lfs diff=lfs merge=lfs -text
69
+ *.mlmodel filter=lfs diff=lfs merge=lfs -text
70
+ *.model filter=lfs diff=lfs merge=lfs -text
71
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
72
+ *.npy filter=lfs diff=lfs merge=lfs -text
73
+ *.npz filter=lfs diff=lfs merge=lfs -text
74
+ *.onnx filter=lfs diff=lfs merge=lfs -text
75
+ *.ot filter=lfs diff=lfs merge=lfs -text
76
+ *.parquet filter=lfs diff=lfs merge=lfs -text
77
+ *.pb filter=lfs diff=lfs merge=lfs -text
78
+ *.pickle filter=lfs diff=lfs merge=lfs -text
79
+ *.pkl filter=lfs diff=lfs merge=lfs -text
80
+ *.pt filter=lfs diff=lfs merge=lfs -text
81
+ *.pth filter=lfs diff=lfs merge=lfs -text
82
+ *.rar filter=lfs diff=lfs merge=lfs -text
83
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
84
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
85
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
86
+ *.tar filter=lfs diff=lfs merge=lfs -text
87
+ *.tflite filter=lfs diff=lfs merge=lfs -text
88
+ *.tgz filter=lfs diff=lfs merge=lfs -text
89
+ *.wasm filter=lfs diff=lfs merge=lfs -text
90
+ *.xz filter=lfs diff=lfs merge=lfs -text
91
+ *.zip filter=lfs diff=lfs merge=lfs -text
92
+ *.zst filter=lfs diff=lfs merge=lfs -text
93
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
94
+ # Audio files - uncompressed
95
+ *.pcm filter=lfs diff=lfs merge=lfs -text
96
+ *.sam filter=lfs diff=lfs merge=lfs -text
97
+ *.raw filter=lfs diff=lfs merge=lfs -text
98
+ # Audio files - compressed
99
+ *.aac filter=lfs diff=lfs merge=lfs -text
100
+ *.flac filter=lfs diff=lfs merge=lfs -text
101
+ *.mp3 filter=lfs diff=lfs merge=lfs -text
102
+ *.ogg filter=lfs diff=lfs merge=lfs -text
103
+ *.wav filter=lfs diff=lfs merge=lfs -text
104
+ # Image files - uncompressed
105
+ *.bmp filter=lfs diff=lfs merge=lfs -text
106
+ *.gif filter=lfs diff=lfs merge=lfs -text
107
+ *.png filter=lfs diff=lfs merge=lfs -text
108
+ *.tiff filter=lfs diff=lfs merge=lfs -text
109
+ # Image files - compressed
110
+ *.jpg filter=lfs diff=lfs merge=lfs -text
111
+ *.jpeg filter=lfs diff=lfs merge=lfs -text
112
+ *.webp filter=lfs diff=lfs merge=lfs -text
113
+ political_data filter=lfs diff=lfs merge=lfs -text
114
+ *.en filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ task_categories:
4
+ - text-classification
5
+ language:
6
+ - en
7
+ tags:
8
+ - politics
9
+ - tweets
10
+ pretty_name: Political Slant Tweets Dataset
11
+ size_categories:
12
+ - 100K<n<1M
13
+ ---
14
+
15
+ ## Political Slant Tweets Dataset
16
+
17
+ Political slant transfer dataset. Contains two classes of political tweets between Democratic and Republican Politicans. This dataset can be used for classification tasks.
18
+
19
+ This is taken directly from:
20
+
21
+ http://tts.speech.cs.cmu.edu/style_models/political_data.tar
22
+ tar -xvf political_data.tar
23
+
24
+ I can't take credit for this dataset; this is largely the work of Prabhumoye et al. This is simply a Wrapper for that dataset. See References below.
25
+
26
+ ## Reference
27
+
28
+ @inproceedings{style_transfer_acl18,
29
+ title={Style Transfer Through Back-Translation},
30
+ author={Prabhumoye, Shrimai and Tsvetkov, Yulia and Salakhutdinov, Ruslan and Black, Alan W},
31
+ year={2018},
32
+ booktitle={Proc. ACL}
33
+ }
34
+
35
+ @inproceedings{rtgender,
36
+ title={{RtGender}: A Corpus for Studying Differential Responses to Gender},
37
+ author={Voigt, Rob and Jurgens, David and Prabhakaran, Vinodkumar and Jurafsky, Dan and Tsvetkov, Yulia},
38
+ year={2018},
39
+ booktitle={Proc. LREC},
40
+ }
political.py ADDED
@@ -0,0 +1,134 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+ from pathlib import Path
3
+ import datasets
4
+ from typing import List
5
+
6
+ STYLE, CLASSIFIER = "style", "classifier"
7
+ _CITATION = """\
8
+ @inproceedings{style_transfer_acl18,
9
+ title={Style Transfer Through Back-Translation},
10
+ author={Prabhumoye, Shrimai and Tsvetkov, Yulia and Salakhutdinov, Ruslan and Black, Alan W},
11
+ year={2018},
12
+ booktitle={Proc. ACL}
13
+ }
14
+ """
15
+
16
+ _DESCRIPTION = """\
17
+ Political slant transfer dataset. Contains two classes of political tweets between Democratic and Republican Politicans. This dataset can be used for classification tasks.
18
+ """
19
+ _HOMEPAGE = "https://github.com/shrimai/Style-Transfer-Through-Back-Translation"
20
+ _LICENSE = "" # could not find.
21
+
22
+
23
+ class PoliticalDataset(datasets.GeneratorBasedBuilder):
24
+ VERSION = datasets.Version("1.0.0")
25
+ DEFAULT_CONFIG_NAME = STYLE
26
+
27
+ BUILDER_CONFIGS = [
28
+ datasets.BuilderConfig(
29
+ name=STYLE,
30
+ version=VERSION,
31
+ description="Political Tweets Dataset, used for Style Transfer tasks.",
32
+ ),
33
+ datasets.BuilderConfig(
34
+ name=CLASSIFIER,
35
+ version=VERSION,
36
+ description="Political Tweets Dataset, Used for classification tasks.",
37
+ ),
38
+ ]
39
+
40
+ def _info(self):
41
+ features = datasets.Features(
42
+ {
43
+ "text": datasets.Value("string"),
44
+ "label": datasets.Value("string"),
45
+ }
46
+ )
47
+ return datasets.DatasetInfo(
48
+ description=_DESCRIPTION,
49
+ features=features,
50
+ supervised_keys=("text", "label"),
51
+ homepage=_HOMEPAGE,
52
+ license=_LICENSE,
53
+ citation=_CITATION,
54
+ )
55
+
56
+ def _split_generators(self, dl_manager) -> List[datasets.SplitGenerator]:
57
+ data_dir = "political_data"
58
+
59
+ splits: List[datasets.SplitGenerator] = []
60
+ if self.config.name == STYLE:
61
+ splits.append(
62
+ datasets.SplitGenerator(
63
+ name=datasets.Split.TRAIN,
64
+ gen_kwargs={
65
+ "filepaths": [
66
+ os.path.join(data_dir, "republican_only.train.en"),
67
+ os.path.join(data_dir, "democratic_only.train.en"),
68
+ ],
69
+ "split": "train",
70
+ },
71
+ )
72
+ )
73
+ else:
74
+ splits.append(
75
+ datasets.SplitGenerator(
76
+ name=datasets.Split.TRAIN,
77
+ gen_kwargs={
78
+ "filepaths": [os.path.join(data_dir, "classtrain.txt")],
79
+ "split": "train",
80
+ },
81
+ )
82
+ )
83
+
84
+ splits += [
85
+ datasets.SplitGenerator(
86
+ name=datasets.Split.VALIDATION,
87
+ gen_kwargs={
88
+ "filepaths": [
89
+ os.path.join(data_dir, "republican_only.dev.en"),
90
+ os.path.join(data_dir, "democratic_only.dev.en"),
91
+ ],
92
+ "split": "dev",
93
+ },
94
+ ),
95
+ datasets.SplitGenerator(
96
+ name=datasets.Split.TEST,
97
+ gen_kwargs={
98
+ "filepaths": [
99
+ os.path.join(data_dir, "republican_only.test.en"),
100
+ os.path.join(data_dir, "democratic_only.test.en"),
101
+ ],
102
+ "split": "test",
103
+ },
104
+ ),
105
+ ]
106
+ return splits
107
+
108
+ def _generate_examples(self, filepaths: List[str], split: str):
109
+ for filepath in filepaths:
110
+ filename = Path(filepath).name
111
+ label = filename.split(".")[0].split("_")[0]
112
+ with open(filepath, encoding="utf-8") as f:
113
+ for key, row in enumerate(f):
114
+ text = row.strip()
115
+ if split != "test":
116
+ # label only exists in train/eval files.
117
+ text = text.split()
118
+ label, text = text[0], text[1:]
119
+ text = " ".join(text)
120
+
121
+ yield (
122
+ key,
123
+ {"text": text, "label": label},
124
+ )
125
+
126
+
127
+ if __name__ == "__main__":
128
+ from tqdm import tqdm
129
+
130
+ dataset = PoliticalDataset(config_name="classifier")
131
+ dataset = dataset.as_streaming_dataset()
132
+ print(dataset)
133
+ for row in tqdm(dataset["train"]):
134
+ row["text"] = "hello"
political_data/classtrain.txt ADDED
The diff for this file is too large to render. See raw diff
 
political_data/democratic_only.dev.en ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:67b0e9358ff8b75a49c761ee3c0f4491fe85cadb486d535d29f814decd77f5b2
3
+ size 186804
political_data/democratic_only.test.en ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:100b382aefbde27b34bb00bbd14fe5c187c16bbdc687557c598111746ab7cc46
3
+ size 2358533
political_data/democratic_only.train.en ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3705adeecd8758859f6b1729ea55cdd327efdbde8b00938d2477b139f45ecc66
3
+ size 25559362
political_data/republican_only.dev.en ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:da2a8ac301562b521dc9922d9635ca030ee5e6dd3846183ae5848266df55ead9
3
+ size 186309
political_data/republican_only.test.en ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15ca942110dcb8d0690ddd8b292d1284bab974087a3bfef5113256ecdb05a38f
3
+ size 2292494
political_data/republican_only.train.en ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c8e609ecabcaee7d1fb96c12a523b9259718b5f4a00a323d24bfdea57457944e
3
+ size 25145176