script 다시 제대로 만듦!!
Browse files- kowikitext.py +16 -7
kowikitext.py
CHANGED
@@ -49,7 +49,7 @@ _LICENSE = "Creative_Commons_Attribution-ShareAlike_3.0_Unported_License"
|
|
49 |
# The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
|
50 |
# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
|
51 |
_URLS = {
|
52 |
-
"20221001": "https://huggingface.co/datasets/heegyu/kowikitext/
|
53 |
}
|
54 |
|
55 |
|
@@ -111,13 +111,22 @@ class KowikitextDataset(datasets.GeneratorBasedBuilder):
|
|
111 |
urls = _URLS[self.config.name]
|
112 |
data_dir = dl_manager.download_and_extract(urls)
|
113 |
|
|
|
114 |
# no split
|
115 |
-
return [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
116 |
|
117 |
# method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
|
118 |
def _generate_examples(self, filepath, split):
|
119 |
-
|
120 |
-
|
121 |
-
|
122 |
-
|
123 |
-
yield key, data
|
|
|
49 |
# The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
|
50 |
# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
|
51 |
_URLS = {
|
52 |
+
"20221001": "https://huggingface.co/datasets/heegyu/kowikitext/resolve/main/kowikitext-20221001.parquet",
|
53 |
}
|
54 |
|
55 |
|
|
|
111 |
urls = _URLS[self.config.name]
|
112 |
data_dir = dl_manager.download_and_extract(urls)
|
113 |
|
114 |
+
|
115 |
# no split
|
116 |
+
return [
|
117 |
+
datasets.SplitGenerator(
|
118 |
+
name=datasets.Split.TRAIN,
|
119 |
+
# These kwargs will be passed to _generate_examples
|
120 |
+
gen_kwargs={
|
121 |
+
"filepath": data_dir,
|
122 |
+
"split": "train",
|
123 |
+
},
|
124 |
+
),
|
125 |
+
]
|
126 |
|
127 |
# method parameters are unpacked from `gen_kwargs` as given in `_split_generators`
|
128 |
def _generate_examples(self, filepath, split):
|
129 |
+
import pandas as pd
|
130 |
+
df = pd.read_parquet(filepath)
|
131 |
+
for key, row in df.iterrows():
|
132 |
+
yield int(key), row.to_dict()
|
|