Kosuke-Yamada
commited on
Commit
•
3a6b34c
1
Parent(s):
9d30a18
modify file
Browse files- ner-wikipedia-dataset.py +33 -5
ner-wikipedia-dataset.py
CHANGED
@@ -4,6 +4,7 @@ from typing import Generator
|
|
4 |
|
5 |
from datasets import (
|
6 |
BuilderConfig,
|
|
|
7 |
DatasetInfo,
|
8 |
DownloadManager,
|
9 |
Features,
|
@@ -30,9 +31,35 @@ _LICENSE = "CC-BY-SA 3.0"
|
|
30 |
_URL = "https://raw.githubusercontent.com/stockmarkteam/ner-wikipedia-dataset/main/ner.json"
|
31 |
|
32 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
33 |
class NerWikipediaDataset(GeneratorBasedBuilder):
|
34 |
BUILDER_CONFIGS = [
|
35 |
-
|
36 |
name="ner-wikipedia-dataset",
|
37 |
version=Version("2.0.0"),
|
38 |
description=_DESCRIPTION,
|
@@ -67,12 +94,13 @@ class NerWikipediaDataset(GeneratorBasedBuilder):
|
|
67 |
with open(dataset_dir, "r", encoding="utf-8") as f:
|
68 |
data = json.load(f)
|
69 |
|
70 |
-
|
71 |
-
|
|
|
72 |
|
73 |
num_data = len(data)
|
74 |
-
num_train_data = int(num_data *
|
75 |
-
num_validation_data = (num_data
|
76 |
train_data = data[:num_train_data]
|
77 |
validation_data = data[
|
78 |
num_train_data : num_train_data + num_validation_data
|
|
|
4 |
|
5 |
from datasets import (
|
6 |
BuilderConfig,
|
7 |
+
DataFilesDict,
|
8 |
DatasetInfo,
|
9 |
DownloadManager,
|
10 |
Features,
|
|
|
31 |
_URL = "https://raw.githubusercontent.com/stockmarkteam/ner-wikipedia-dataset/main/ner.json"
|
32 |
|
33 |
|
34 |
+
class NerWikipediaDatasetConfig(BuilderConfig):
|
35 |
+
def __init__(
|
36 |
+
self,
|
37 |
+
name: str = "default",
|
38 |
+
version: Version | str | None = Version("0.0.0"),
|
39 |
+
data_dir: str | None = None,
|
40 |
+
data_files: DataFilesDict | None = None,
|
41 |
+
description: str | None = None,
|
42 |
+
shuffle: bool = True,
|
43 |
+
seed: int = 42,
|
44 |
+
train_ratio: float = 0.8,
|
45 |
+
validation_ratio: float = 0.1,
|
46 |
+
) -> None:
|
47 |
+
super().__init__(
|
48 |
+
name=name,
|
49 |
+
version=version,
|
50 |
+
data_dir=data_dir,
|
51 |
+
data_files=data_files,
|
52 |
+
description=description,
|
53 |
+
)
|
54 |
+
self.shuffle = shuffle
|
55 |
+
self.seed = seed
|
56 |
+
self.train_ratio = train_ratio
|
57 |
+
self.validation_ratio = validation_ratio
|
58 |
+
|
59 |
+
|
60 |
class NerWikipediaDataset(GeneratorBasedBuilder):
|
61 |
BUILDER_CONFIGS = [
|
62 |
+
NerWikipediaDatasetConfig(
|
63 |
name="ner-wikipedia-dataset",
|
64 |
version=Version("2.0.0"),
|
65 |
description=_DESCRIPTION,
|
|
|
94 |
with open(dataset_dir, "r", encoding="utf-8") as f:
|
95 |
data = json.load(f)
|
96 |
|
97 |
+
if self.config.shuffle == True:
|
98 |
+
random.seed(self.config.seed)
|
99 |
+
random.shuffle(data)
|
100 |
|
101 |
num_data = len(data)
|
102 |
+
num_train_data = int(num_data * self.config.train_ratio)
|
103 |
+
num_validation_data = int(num_data * self.config.validation_ratio)
|
104 |
train_data = data[:num_train_data]
|
105 |
validation_data = data[
|
106 |
num_train_data : num_train_data + num_validation_data
|