qgyd2021 commited on
Commit
4eff5ab
1 Parent(s): 12b173a

[update]add ocnli

Browse files
README.md CHANGED
@@ -21,7 +21,9 @@ size_categories:
21
  | BUSTM | 汉语 | [BUSTM](https://tianchi.aliyun.com/competition/entrance/531851/information); [BUSTM](https://github.com/xiaobu-coai/BUSTM) | 总样本数为:177173,其中,匹配样本个数为:54805,不匹配样本个数为:122368 | 小布助手对话短文本语义匹配比赛数据集 | [BUSTM](https://github.com/CLUEbenchmark/FewCLUE/tree/main/datasets/bustm) |
22
  | CHIP2019 | 汉语 | [CHIP2019](https://www.biendata.xyz/competition/chip2019/) | 2万 | 平安医疗科技疾病问答迁移学习比赛数据集 | |
23
  | COVID-19 | 汉语 | [COVID-19](https://tianchi.aliyun.com/competition/entrance/231776/information) | | 天池新冠疫情相似句对判定大赛 | [COVID-19](https://gitee.com/liangzongchang/COVID-19-sentence-pair/) |
24
- | Chinese-MNLI | 汉语 | [Chinese-MNLI](https://github.com/pluto-junzeng/CNSD) | TRAIN: 390K, VALID: 12K, TEST: 13K | 通过翻译加部分人工修正的方法,从英文原数据集生成(原数据是:蕴含,中性,冲突,的句子推理数据集,已转换为句子对。) | |
 
 
25
 
26
 
27
  <details>
 
21
  | BUSTM | 汉语 | [BUSTM](https://tianchi.aliyun.com/competition/entrance/531851/information); [BUSTM](https://github.com/xiaobu-coai/BUSTM) | 总样本数为:177173,其中,匹配样本个数为:54805,不匹配样本个数为:122368 | 小布助手对话短文本语义匹配比赛数据集 | [BUSTM](https://github.com/CLUEbenchmark/FewCLUE/tree/main/datasets/bustm) |
22
  | CHIP2019 | 汉语 | [CHIP2019](https://www.biendata.xyz/competition/chip2019/) | 2万 | 平安医疗科技疾病问答迁移学习比赛数据集 | |
23
  | COVID-19 | 汉语 | [COVID-19](https://tianchi.aliyun.com/competition/entrance/231776/information) | | 天池新冠疫情相似句对判定大赛 | [COVID-19](https://gitee.com/liangzongchang/COVID-19-sentence-pair/) |
24
+ | Chinese-MNLI | 汉语 | [Chinese-MNLI](https://github.com/pluto-junzeng/CNSD) | TRAIN: 390K, VALID: 12K, TEST: 13K | 通过翻译加部分人工修正的方法,从英文原数据集生成(原数据是:蕴含,中性,冲突,的句子推理数据集,已转换为句子对)。 | |
25
+ | Chinese-SNLI | 汉语 | [Chinese-SNLI](https://github.com/pluto-junzeng/CNSD) | TRAIN: 550K, VALID: 10K, TEST: 10K | 通过翻译加部分人工修正的方法,从英文原数据集生成(原数据是:蕴含,中性,冲突,的句子推理数据集,已转换为句子对)。 | |
26
+ | OCNLI | 汉语 | [OCNLI](https://github.com/CLUEbenchmark/OCNLI) | TRAIN: 50K, VALID: 3K, TEST: 3K | 原生中文自然语言推理数据集,是第一个非翻译的、使用原生汉语的大型中文自然语言推理数据集。 | |
27
 
28
 
29
  <details>
data/chinese_mnli.jsonl CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:994e50bf6964ca0d6073b31fae10af95c2037ebbba434a864591b79c64757045
3
- size 117184478
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f0b87b6ebc3b8124ca4f1cbb0e5055e4941c0f8f93014248f93d82cbf617a585
3
+ size 117078623
data/chinese_snli.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0616f76d77f83a4bbc4ff0b0cbea1a018dbdcf8c7dc94538a062284d9626704f
3
+ size 125893606
data/ocnli.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12b5f3a0e082d09f29d646024f73c870906c29adc81cdbd9c6c519fe276c3f33
3
+ size 12092258
examples/preprocess/process_chinese_mnli.py CHANGED
@@ -45,16 +45,22 @@ def main():
45
  sentence2 = row["sentence2"]
46
  gold_label = row["gold_label"]
47
 
 
 
 
 
 
 
 
 
 
48
  if name == "cnsd_multil_train.jsonl":
49
- label = "1" if gold_label in ("entailment",) else "0"
50
  flag = "train"
51
  category = "cnsd_multil_train"
52
  elif name == "cnsd_multil_dev_matched.jsonl":
53
- label = "1" if gold_label in ("entailment",) else "0"
54
  flag = "validation"
55
  category = "cnsd_multil_dev_matched"
56
  elif name == "cnsd_multil_dev_mismatched.jsonl":
57
- label = "1" if gold_label in ("entailment",) else "0"
58
  flag = "validation"
59
  category = "cnsd_multil_dev_mismatched"
60
  else:
 
45
  sentence2 = row["sentence2"]
46
  gold_label = row["gold_label"]
47
 
48
+ if gold_label not in ("entailment", "neutral", "contradiction", "-"):
49
+ print(gold_label)
50
+ raise AssertionError
51
+
52
+ if gold_label in ("-",):
53
+ continue
54
+
55
+ label = "1" if gold_label in ("entailment",) else "0"
56
+
57
  if name == "cnsd_multil_train.jsonl":
 
58
  flag = "train"
59
  category = "cnsd_multil_train"
60
  elif name == "cnsd_multil_dev_matched.jsonl":
 
61
  flag = "validation"
62
  category = "cnsd_multil_dev_matched"
63
  elif name == "cnsd_multil_dev_mismatched.jsonl":
 
64
  flag = "validation"
65
  category = "cnsd_multil_dev_mismatched"
66
  else:
examples/preprocess/process_chinese_snli.py ADDED
@@ -0,0 +1,86 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/python3
2
+ # -*- coding: utf-8 -*-
3
+ import argparse
4
+ import json
5
+ import os
6
+ from pathlib import Path
7
+ import sys
8
+
9
+ pwd = os.path.abspath(os.path.dirname(__file__))
10
+ sys.path.append(os.path.join(pwd, '../../'))
11
+
12
+ from datasets import load_dataset
13
+ from tqdm import tqdm
14
+
15
+ from project_settings import project_path
16
+
17
+
18
+ def get_args():
19
+ parser = argparse.ArgumentParser()
20
+
21
+ parser.add_argument("--data_dir", default="./data/chinese_snli", type=str)
22
+
23
+ parser.add_argument(
24
+ "--output_file",
25
+ default=(project_path / "data/chinese_snli.jsonl"),
26
+ type=str
27
+ )
28
+
29
+ args = parser.parse_args()
30
+ return args
31
+
32
+
33
+ def main():
34
+ args = get_args()
35
+
36
+ data_dir = Path(args.data_dir)
37
+
38
+ with open(args.output_file, "w", encoding="utf-8") as fout:
39
+ for name in ["cnsd_snli_v1.0.train.jsonl", "cnsd_snli_v1.0.dev.jsonl", "cnsd_snli_v1.0.test.jsonl"]:
40
+ filename = data_dir / name
41
+ with open(filename, "r", encoding="utf-8") as fin:
42
+ for row in fin:
43
+ row = json.loads(row)
44
+ sentence1 = row["sentence1"]
45
+ sentence2 = row["sentence2"]
46
+ gold_label = row["gold_label"]
47
+
48
+ if gold_label not in ("entailment", "neutral", "contradiction", "-"):
49
+ print(gold_label)
50
+ raise AssertionError
51
+
52
+ if gold_label in ("-",):
53
+ continue
54
+
55
+ label = "1" if gold_label in ("entailment",) else "0"
56
+ category = gold_label
57
+
58
+ if name == "cnsd_snli_v1.0.train.jsonl":
59
+ flag = "train"
60
+ elif name == "cnsd_snli_v1.0.dev.jsonl":
61
+ flag = "validation"
62
+ elif name == "cnsd_snli_v1.0.test.jsonl":
63
+ flag = "test"
64
+ else:
65
+ raise AssertionError
66
+
67
+ if label not in ("0", "1", None):
68
+ raise AssertionError
69
+
70
+ row = {
71
+ "sentence1": sentence1,
72
+ "sentence2": sentence2,
73
+ "label": label,
74
+ "category": category,
75
+ "data_source": "Chinese-SNLI",
76
+ "split": flag
77
+ }
78
+
79
+ row = json.dumps(row, ensure_ascii=False)
80
+ fout.write("{}\n".format(row))
81
+
82
+ return
83
+
84
+
85
+ if __name__ == '__main__':
86
+ main()
examples/preprocess/process_ocnli.py ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/python3
2
+ # -*- coding: utf-8 -*-
3
+ import argparse
4
+ import json
5
+ import os
6
+ from pathlib import Path
7
+ import sys
8
+
9
+ pwd = os.path.abspath(os.path.dirname(__file__))
10
+ sys.path.append(os.path.join(pwd, '../../'))
11
+
12
+ from datasets import load_dataset
13
+ from tqdm import tqdm
14
+
15
+ from project_settings import project_path
16
+
17
+
18
+ def get_args():
19
+ parser = argparse.ArgumentParser()
20
+
21
+ parser.add_argument("--data_dir", default="./data/ocnli/data/ocnli", type=str)
22
+
23
+ parser.add_argument(
24
+ "--output_file",
25
+ default=(project_path / "data/ocnli.jsonl"),
26
+ type=str
27
+ )
28
+
29
+ args = parser.parse_args()
30
+ return args
31
+
32
+
33
+ def main():
34
+ args = get_args()
35
+
36
+ data_dir = Path(args.data_dir)
37
+
38
+ with open(args.output_file, "w", encoding="utf-8") as fout:
39
+ for name in ["train.50k.json", "dev.json", "test.json"]:
40
+ filename = data_dir / name
41
+ with open(filename, "r", encoding="utf-8") as fin:
42
+ for row in fin:
43
+ row = json.loads(row)
44
+ sentence1 = row["sentence1"]
45
+ sentence2 = row["sentence2"]
46
+ label = row.get("label", None)
47
+ category = row.get("genre", None)
48
+
49
+ if label not in ("entailment", "neutral", "contradiction", "-", None):
50
+ print(label)
51
+ raise AssertionError
52
+
53
+ if label in ("-",):
54
+ continue
55
+
56
+ if label is not None:
57
+ label = "1" if label in ("entailment",) else "0"
58
+
59
+ if name == "train.50k.json":
60
+ flag = "train"
61
+ elif name == "dev.json":
62
+ flag = "validation"
63
+ elif name == "test.json":
64
+ flag = "test"
65
+ else:
66
+ raise AssertionError
67
+
68
+ row = {
69
+ "sentence1": sentence1,
70
+ "sentence2": sentence2,
71
+ "label": label,
72
+ "category": category,
73
+ "data_source": "OCNLI",
74
+ "split": flag
75
+ }
76
+
77
+ row = json.dumps(row, ensure_ascii=False)
78
+ fout.write("{}\n".format(row))
79
+
80
+ return
81
+
82
+
83
+ if __name__ == '__main__':
84
+ main()
sentence_pair.py CHANGED
@@ -13,11 +13,13 @@ _URLS = {
13
  "bustm": "data/bustm.jsonl",
14
  "ccks2018_task3": "data/ccks2018_task3.jsonl",
15
  "chinese_mnli": "data/chinese_mnli.jsonl",
 
16
  "chinese_sts": "data/chinese_sts.jsonl",
17
  "chip2019": "data/chip2019.jsonl",
18
  "covid_19": "data/covid_19.jsonl",
19
  "diac2019": "data/diac2019.jsonl",
20
  "lcqmc": "data/lcqmc.jsonl",
 
21
 
22
  }
23
 
 
13
  "bustm": "data/bustm.jsonl",
14
  "ccks2018_task3": "data/ccks2018_task3.jsonl",
15
  "chinese_mnli": "data/chinese_mnli.jsonl",
16
+ "chinese_snli": "data/chinese_snli.jsonl",
17
  "chinese_sts": "data/chinese_sts.jsonl",
18
  "chip2019": "data/chip2019.jsonl",
19
  "covid_19": "data/covid_19.jsonl",
20
  "diac2019": "data/diac2019.jsonl",
21
  "lcqmc": "data/lcqmc.jsonl",
22
+ "ocnli": "data/ocnli.jsonl",
23
 
24
  }
25