Datasets:
parquet-converter
commited on
Commit
•
73a4950
1
Parent(s):
d83da96
Update parquet files
Browse files- README.md +0 -130
- amharic/eval.zip → afaanoromoo/afriberta-corpus-test.parquet +2 -2
- afaanoromoo/{train.zip → afriberta-corpus-train.parquet} +2 -2
- afriberta-corpus.py +0 -113
- afaanoromoo/eval.zip → amharic/afriberta-corpus-test.parquet +2 -2
- amharic/afriberta-corpus-train.parquet +3 -0
- dataset_infos.json +0 -1
- gahuza/{eval.zip → afriberta-corpus-test.parquet} +2 -2
- amharic/train.zip → gahuza/afriberta-corpus-train.parquet +2 -2
- gahuza/train.zip → hausa/afriberta-corpus-test.parquet +2 -2
- hausa/afriberta-corpus-train.parquet +3 -0
- hausa/eval.zip +0 -3
- hausa/train.zip +0 -3
- igbo/afriberta-corpus-test.parquet +3 -0
- igbo/afriberta-corpus-train.parquet +3 -0
- igbo/eval.zip +0 -3
- igbo/train.zip +0 -3
- pidgin/afriberta-corpus-test.parquet +3 -0
- pidgin/afriberta-corpus-train.parquet +3 -0
- pidgin/eval.zip +0 -3
- pidgin/train.zip +0 -3
- somali/afriberta-corpus-test.parquet +3 -0
- somali/afriberta-corpus-train.parquet +3 -0
- somali/eval.zip +0 -3
- somali/train.zip +0 -3
- swahili/afriberta-corpus-test.parquet +3 -0
- swahili/afriberta-corpus-train.parquet +3 -0
- swahili/eval.zip +0 -3
- swahili/train.zip +0 -3
- tigrinya/afriberta-corpus-test.parquet +3 -0
- tigrinya/afriberta-corpus-train.parquet +3 -0
- tigrinya/eval.zip +0 -3
- tigrinya/train.zip +0 -3
- yoruba/afriberta-corpus-test.parquet +3 -0
- yoruba/afriberta-corpus-train.parquet +3 -0
- yoruba/eval.zip +0 -3
- yoruba/train.zip +0 -3
README.md
DELETED
@@ -1,130 +0,0 @@
|
|
1 |
-
---
|
2 |
-
language:
|
3 |
-
- om
|
4 |
-
- am
|
5 |
-
- rw
|
6 |
-
- rn
|
7 |
-
- ha
|
8 |
-
- ig
|
9 |
-
- pcm
|
10 |
-
- so
|
11 |
-
- sw
|
12 |
-
- ti
|
13 |
-
- yo
|
14 |
-
- multilingual
|
15 |
-
|
16 |
-
license: apache-2.0
|
17 |
-
|
18 |
-
task_categories:
|
19 |
-
- text-generation
|
20 |
-
task_ids:
|
21 |
-
- language-modeling
|
22 |
-
---
|
23 |
-
|
24 |
-
# Dataset Card for AfriBERTa's Corpus
|
25 |
-
|
26 |
-
## Table of Contents
|
27 |
-
- [Dataset Description](#dataset-description)
|
28 |
-
- [Dataset Summary](#dataset-summary)
|
29 |
-
- [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
|
30 |
-
- [Languages](#languages)
|
31 |
-
- [Loading Dataset](#loading-dataset)
|
32 |
-
- [Dataset Structure](#dataset-structure)
|
33 |
-
- [Data Instances](#data-instances)
|
34 |
-
- [Data Fields](#data-fields)
|
35 |
-
- [Data Splits](#data-splits)
|
36 |
-
- [Considerations for Using the Data](#considerations-for-using-the-data)
|
37 |
-
- [Discussion of Biases](#discussion-of-biases)
|
38 |
-
- [Additional Information](#additional-information)
|
39 |
-
- [Citation Information](#citation-information)
|
40 |
-
- [Contributions](#contributions)
|
41 |
-
|
42 |
-
## Dataset Description
|
43 |
-
|
44 |
-
### Dataset Summary
|
45 |
-
This is the corpus on which AfriBERTa was trained on.
|
46 |
-
The dataset is mostly from the BBC news website, but some languages also have data from Common Crawl.
|
47 |
-
|
48 |
-
- **Homepage:** https://github.com/keleog/afriberta
|
49 |
-
- **Models:**
|
50 |
-
- https://huggingface.co/castorini/afriberta_small
|
51 |
-
- https://huggingface.co/castorini/afriberta_base
|
52 |
-
- https://huggingface.co/castorini/afriberta_large
|
53 |
-
- **Paper:** https://aclanthology.org/2021.mrl-1.11/
|
54 |
-
- **Point of Contact:** kelechi.ogueji@uwaterloo.ca
|
55 |
-
|
56 |
-
|
57 |
-
### Supported Tasks and Leaderboards
|
58 |
-
The AfriBERTa corpus was mostly intended to pre-train language models.
|
59 |
-
|
60 |
-
### Languages
|
61 |
-
```
|
62 |
-
afaanoromoo
|
63 |
-
amharic
|
64 |
-
gahuza
|
65 |
-
hausa
|
66 |
-
igbo
|
67 |
-
pidgin
|
68 |
-
somali
|
69 |
-
swahili
|
70 |
-
tigrinya
|
71 |
-
yoruba
|
72 |
-
```
|
73 |
-
### Loading Dataset
|
74 |
-
An example to load the train split of the Somali corpus:
|
75 |
-
```
|
76 |
-
dataset = load_dataset("castorini/afriberta-corpus", "somali", split="train")
|
77 |
-
```
|
78 |
-
|
79 |
-
An example to load the test split of the Pidgin corpus:
|
80 |
-
```
|
81 |
-
dataset = load_dataset("castorini/afriberta-corpus", "pidgin", split="test")
|
82 |
-
```
|
83 |
-
|
84 |
-
## Dataset Structure
|
85 |
-
|
86 |
-
### Data Instances
|
87 |
-
Each data point is a line of text.
|
88 |
-
An example from the `igbo` dataset:
|
89 |
-
```
|
90 |
-
{"id": "6", "text": "Ngwá ọrụ na-echebe ma na-ebuli gị na kọmputa."}
|
91 |
-
```
|
92 |
-
|
93 |
-
### Data Fields
|
94 |
-
|
95 |
-
The data fields are:
|
96 |
-
|
97 |
-
- id: id of the example
|
98 |
-
- text: content as a string
|
99 |
-
|
100 |
-
### Data Splits
|
101 |
-
Each language has a train and test split, with varying sizes.
|
102 |
-
|
103 |
-
## Considerations for Using the Data
|
104 |
-
|
105 |
-
### Discussion of Biases
|
106 |
-
Since majority of the data is obtained from the BBC's news website, models trained on this dataset are likely going to
|
107 |
-
be biased towards the news domain.
|
108 |
-
|
109 |
-
Also, since some of the data is obtained from Common Crawl, care should be taken (especially for text generation models) since personal and sensitive information might be present.
|
110 |
-
|
111 |
-
## Additional Information
|
112 |
-
### Citation Information
|
113 |
-
```
|
114 |
-
@inproceedings{ogueji-etal-2021-small,
|
115 |
-
title = "Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages",
|
116 |
-
author = "Ogueji, Kelechi and
|
117 |
-
Zhu, Yuxin and
|
118 |
-
Lin, Jimmy",
|
119 |
-
booktitle = "Proceedings of the 1st Workshop on Multilingual Representation Learning",
|
120 |
-
month = nov,
|
121 |
-
year = "2021",
|
122 |
-
address = "Punta Cana, Dominican Republic",
|
123 |
-
publisher = "Association for Computational Linguistics",
|
124 |
-
url = "https://aclanthology.org/2021.mrl-1.11",
|
125 |
-
pages = "116--126",
|
126 |
-
}
|
127 |
-
```
|
128 |
-
|
129 |
-
### Contributions
|
130 |
-
Thanks to [Kelechi Ogueji](https://github.com/keleog) for adding this dataset.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
amharic/eval.zip → afaanoromoo/afriberta-corpus-test.parquet
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:441a6fc1542fd23c3b17f731653f7c007c11ebedaddc815cd703de3204ac651c
|
3 |
+
size 2849238
|
afaanoromoo/{train.zip → afriberta-corpus-train.parquet}
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:18f8389d2ade9a006f7690c110d1860db5664fca96cf1ae027cb547b4e0bfea5
|
3 |
+
size 38946179
|
afriberta-corpus.py
DELETED
@@ -1,113 +0,0 @@
|
|
1 |
-
# coding=utf-8
|
2 |
-
# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
|
3 |
-
#
|
4 |
-
# Licensed under the Apache License, Version 2.0 (the 'License');
|
5 |
-
# you may not use this file except in compliance with the License.
|
6 |
-
# You may obtain a copy of the License at
|
7 |
-
#
|
8 |
-
# http://www.apache.org/licenses/LICENSE-2.0
|
9 |
-
#
|
10 |
-
# Unless required by applicable law or agreed to in writing, software
|
11 |
-
# distributed under the License is distributed on an 'AS IS' BASIS,
|
12 |
-
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
13 |
-
# See the License for the specific language governing permissions and
|
14 |
-
# limitations under the License.
|
15 |
-
|
16 |
-
# Lint as: python3
|
17 |
-
|
18 |
-
import os
|
19 |
-
import datasets
|
20 |
-
|
21 |
-
_DESCRIPTION = """\
|
22 |
-
Corpus used for training AfriBERTa models
|
23 |
-
"""
|
24 |
-
_CITATION = """\
|
25 |
-
@inproceedings{ogueji-etal-2021-small,
|
26 |
-
title = "Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages",
|
27 |
-
author = "Ogueji, Kelechi and
|
28 |
-
Zhu, Yuxin and
|
29 |
-
Lin, Jimmy",
|
30 |
-
booktitle = "Proceedings of the 1st Workshop on Multilingual Representation Learning",
|
31 |
-
month = nov,
|
32 |
-
year = "2021",
|
33 |
-
address = "Punta Cana, Dominican Republic",
|
34 |
-
publisher = "Association for Computational Linguistics",
|
35 |
-
url = "https://aclanthology.org/2021.mrl-1.11",
|
36 |
-
pages = "116--126",
|
37 |
-
}
|
38 |
-
"""
|
39 |
-
_HOMEPAGE_URL = "https://github.com/keleog/afriberta"
|
40 |
-
_VERSION = "1.0.0"
|
41 |
-
_LANGUAGES = [
|
42 |
-
"afaanoromoo",
|
43 |
-
"amharic",
|
44 |
-
"gahuza",
|
45 |
-
"hausa",
|
46 |
-
"igbo",
|
47 |
-
"pidgin",
|
48 |
-
"somali",
|
49 |
-
"swahili",
|
50 |
-
"tigrinya",
|
51 |
-
"yoruba"]
|
52 |
-
|
53 |
-
_DATASET_URLS = {
|
54 |
-
language: {
|
55 |
-
"train": f"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/{language}/train.zip",
|
56 |
-
"test": f"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/{language}/eval.zip",
|
57 |
-
} for language in _LANGUAGES
|
58 |
-
}
|
59 |
-
|
60 |
-
class AfribertaCorpus(datasets.GeneratorBasedBuilder):
|
61 |
-
BUILDER_CONFIGS = [
|
62 |
-
datasets.BuilderConfig(
|
63 |
-
version=datasets.Version(_VERSION),
|
64 |
-
name=language,
|
65 |
-
description=f"AfriBERTa corpus for {language}."
|
66 |
-
) for language in _LANGUAGES
|
67 |
-
]
|
68 |
-
|
69 |
-
def _info(self):
|
70 |
-
return datasets.DatasetInfo(
|
71 |
-
description=_DESCRIPTION,
|
72 |
-
features=datasets.Features(
|
73 |
-
{
|
74 |
-
"id": datasets.Value("string"),
|
75 |
-
"text": datasets.Value("string"),
|
76 |
-
},
|
77 |
-
),
|
78 |
-
supervised_keys=None,
|
79 |
-
homepage=_HOMEPAGE_URL,
|
80 |
-
citation=_CITATION,
|
81 |
-
)
|
82 |
-
|
83 |
-
def _split_generators(self, dl_manager):
|
84 |
-
language = self.config.name
|
85 |
-
downloaded_files = dl_manager.download_and_extract(_DATASET_URLS[language])
|
86 |
-
|
87 |
-
splits = [
|
88 |
-
datasets.SplitGenerator(
|
89 |
-
name=datasets.Split.TRAIN,
|
90 |
-
gen_kwargs={
|
91 |
-
"file_path": os.path.join(downloaded_files["train"], "train.txt"),
|
92 |
-
},
|
93 |
-
),
|
94 |
-
datasets.SplitGenerator(
|
95 |
-
name=datasets.Split.TEST,
|
96 |
-
gen_kwargs={
|
97 |
-
"file_path": os.path.join(downloaded_files["test"], "eval.txt"),
|
98 |
-
},
|
99 |
-
),
|
100 |
-
]
|
101 |
-
return splits
|
102 |
-
|
103 |
-
def _generate_examples(self, file_path):
|
104 |
-
with open(file_path, encoding="utf-8") as f:
|
105 |
-
for sentence_counter, line in enumerate(f):
|
106 |
-
result = (
|
107 |
-
sentence_counter,
|
108 |
-
{
|
109 |
-
"id": str(sentence_counter),
|
110 |
-
"text": line,
|
111 |
-
},
|
112 |
-
)
|
113 |
-
yield result
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
afaanoromoo/eval.zip → amharic/afriberta-corpus-test.parquet
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:45bb2d70c39c85f100894ac31c02e93b265d9c94ff5f4bff00d5b36ed225ad66
|
3 |
+
size 7351975
|
amharic/afriberta-corpus-train.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:af08443e1f3874841863f97db3de5bec0ad375add1be67f79e6f2305c07d7961
|
3 |
+
size 127357990
|
dataset_infos.json
DELETED
@@ -1 +0,0 @@
|
|
1 |
-
{"afaanoromoo": {"description": "Corpus used for training AfriBERTa models\n", "citation": "@inproceedings{ogueji-etal-2021-small,\n title = \"Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages\",\n author = \"Ogueji, Kelechi and\n Zhu, Yuxin and\n Lin, Jimmy\",\n booktitle = \"Proceedings of the 1st Workshop on Multilingual Representation Learning\",\n month = nov,\n year = \"2021\",\n address = \"Punta Cana, Dominican Republic\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/2021.mrl-1.11\",\n pages = \"116--126\",\n}\n", "homepage": "https://github.com/keleog/afriberta", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "afriberta_corpus", "config_name": "afaanoromoo", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 58836661, "num_examples": 410841, "dataset_name": "afriberta_corpus"}, "test": {"name": "test", "num_bytes": 4261710, "num_examples": 30000, "dataset_name": "afriberta_corpus"}}, "download_checksums": {"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/afaanoromoo/train.zip": {"num_bytes": 20842874, "checksum": "06061756970746d77188461866c859523170731395217978372791bf0ccc22f5"}, "https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/afaanoromoo/eval.zip": {"num_bytes": 1525285, "checksum": "0dbf7a83bcf153a49ad9dea41e478c9526fcdf781e8f7a01168279f24cdc78b5"}}, "download_size": 22368159, "post_processing_size": null, "dataset_size": 63098371, "size_in_bytes": 85466530}, "amharic": {"description": "Corpus used for training AfriBERTa models\n", "citation": "@inproceedings{ogueji-etal-2021-small,\n title = \"Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages\",\n author = \"Ogueji, Kelechi and\n Zhu, Yuxin and\n Lin, Jimmy\",\n booktitle = \"Proceedings of the 1st Workshop on Multilingual Representation Learning\",\n month = nov,\n year = \"2021\",\n address = \"Punta Cana, Dominican Republic\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/2021.mrl-1.11\",\n pages = \"116--126\",\n}\n", "homepage": "https://github.com/keleog/afriberta", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "afriberta_corpus", "config_name": "amharic", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 230146342, "num_examples": 525025, "dataset_name": "afriberta_corpus"}, "test": {"name": "test", "num_bytes": 13260423, "num_examples": 30000, "dataset_name": "afriberta_corpus"}}, "download_checksums": {"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/amharic/train.zip": {"num_bytes": 72059056, "checksum": "a2902804942792ebead91aaec29b118e60c134fddd2327ddf4cc344b77e003d0"}, "https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/amharic/eval.zip": {"num_bytes": 4160661, "checksum": "d05bcb9a90d3862f08f5c0a376a9a268aaa4fbac82c7fec6cd5e72ba28c2d94c"}}, "download_size": 76219717, "post_processing_size": null, "dataset_size": 243406765, "size_in_bytes": 319626482}, "gahuza": {"description": "Corpus used for training AfriBERTa models\n", "citation": "@inproceedings{ogueji-etal-2021-small,\n title = \"Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages\",\n author = \"Ogueji, Kelechi and\n Zhu, Yuxin and\n Lin, Jimmy\",\n booktitle = \"Proceedings of the 1st Workshop on Multilingual Representation Learning\",\n month = nov,\n year = \"2021\",\n address = \"Punta Cana, Dominican Republic\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/2021.mrl-1.11\",\n pages = \"116--126\",\n}\n", "homepage": "https://github.com/keleog/afriberta", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "afriberta_corpus", "config_name": "gahuza", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 28004387, "num_examples": 131953, "dataset_name": "afriberta_corpus"}, "test": {"name": "test", "num_bytes": 3178925, "num_examples": 15000, "dataset_name": "afriberta_corpus"}}, "download_checksums": {"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/gahuza/train.zip": {"num_bytes": 8780402, "checksum": "74e7bd965966b268ef849f825465af43a740b653e3115f13960459db9c591a83"}, "https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/gahuza/eval.zip": {"num_bytes": 1002841, "checksum": "1a9fc0822e3076a774ecdb6513b1298876ea836eb8b348a6f857559a72d95361"}}, "download_size": 9783243, "post_processing_size": null, "dataset_size": 31183312, "size_in_bytes": 40966555}, "hausa": {"description": "Corpus used for training AfriBERTa models\n", "citation": "@inproceedings{ogueji-etal-2021-small,\n title = \"Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages\",\n author = \"Ogueji, Kelechi and\n Zhu, Yuxin and\n Lin, Jimmy\",\n booktitle = \"Proceedings of the 1st Workshop on Multilingual Representation Learning\",\n month = nov,\n year = \"2021\",\n address = \"Punta Cana, Dominican Republic\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/2021.mrl-1.11\",\n pages = \"116--126\",\n}\n", "homepage": "https://github.com/keleog/afriberta", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "afriberta_corpus", "config_name": "hausa", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 175106385, "num_examples": 1282997, "dataset_name": "afriberta_corpus"}, "test": {"name": "test", "num_bytes": 13515044, "num_examples": 100000, "dataset_name": "afriberta_corpus"}}, "download_checksums": {"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/hausa/train.zip": {"num_bytes": 63388604, "checksum": "7c6fa68b7714fe990c94d27b7825282905d2cb3baaddf5d4217b41eae0ffc605"}, "https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/hausa/eval.zip": {"num_bytes": 4946084, "checksum": "eced39cebce201f93cc0f36fe61437b0e1d11dd67628f17c5c19a057fb45b59a"}}, "download_size": 68334688, "post_processing_size": null, "dataset_size": 188621429, "size_in_bytes": 256956117}, "igbo": {"description": "Corpus used for training AfriBERTa models\n", "citation": "@inproceedings{ogueji-etal-2021-small,\n title = \"Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages\",\n author = \"Ogueji, Kelechi and\n Zhu, Yuxin and\n Lin, Jimmy\",\n booktitle = \"Proceedings of the 1st Workshop on Multilingual Representation Learning\",\n month = nov,\n year = \"2021\",\n address = \"Punta Cana, Dominican Republic\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/2021.mrl-1.11\",\n pages = \"116--126\",\n}\n", "homepage": "https://github.com/keleog/afriberta", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "afriberta_corpus", "config_name": "igbo", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 47908238, "num_examples": 337082, "dataset_name": "afriberta_corpus"}, "test": {"name": "test", "num_bytes": 4241199, "num_examples": 30000, "dataset_name": "afriberta_corpus"}}, "download_checksums": {"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/igbo/train.zip": {"num_bytes": 15695174, "checksum": "0b42e54d602fdc47bbe736151222cf4e648340abebc49e0ab8143ef181a591de"}, "https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/igbo/eval.zip": {"num_bytes": 1398211, "checksum": "4a43a72c19f136f3893ad7cde1f8588ae559358028295dec1f0267d561d503ba"}}, "download_size": 17093385, "post_processing_size": null, "dataset_size": 52149437, "size_in_bytes": 69242822}, "pidgin": {"description": "Corpus used for training AfriBERTa models\n", "citation": "@inproceedings{ogueji-etal-2021-small,\n title = \"Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages\",\n author = \"Ogueji, Kelechi and\n Zhu, Yuxin and\n Lin, Jimmy\",\n booktitle = \"Proceedings of the 1st Workshop on Multilingual Representation Learning\",\n month = nov,\n year = \"2021\",\n address = \"Punta Cana, Dominican Republic\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/2021.mrl-1.11\",\n pages = \"116--126\",\n}\n", "homepage": "https://github.com/keleog/afriberta", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "afriberta_corpus", "config_name": "pidgin", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 52355718, "num_examples": 161843, "dataset_name": "afriberta_corpus"}, "test": {"name": "test", "num_bytes": 4831799, "num_examples": 15000, "dataset_name": "afriberta_corpus"}}, "download_checksums": {"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/pidgin/train.zip": {"num_bytes": 13121362, "checksum": "6d52b73d4034f8cd88822b64b8bf3592fae54d89d58c13774b334b38c64034d8"}, "https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/pidgin/eval.zip": {"num_bytes": 1220740, "checksum": "495119efdf722d10aa69a404f0b2b05d8379318531fc0d018acfd0d9aef9b467"}}, "download_size": 14342102, "post_processing_size": null, "dataset_size": 57187517, "size_in_bytes": 71529619}, "somali": {"description": "Corpus used for training AfriBERTa models\n", "citation": "@inproceedings{ogueji-etal-2021-small,\n title = \"Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages\",\n author = \"Ogueji, Kelechi and\n Zhu, Yuxin and\n Lin, Jimmy\",\n booktitle = \"Proceedings of the 1st Workshop on Multilingual Representation Learning\",\n month = nov,\n year = \"2021\",\n address = \"Punta Cana, Dominican Republic\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/2021.mrl-1.11\",\n pages = \"116--126\",\n}\n", "homepage": "https://github.com/keleog/afriberta", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "afriberta_corpus", "config_name": "somali", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 191977457, "num_examples": 995044, "dataset_name": "afriberta_corpus"}, "test": {"name": "test", "num_bytes": 19235708, "num_examples": 100000, "dataset_name": "afriberta_corpus"}}, "download_checksums": {"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/somali/train.zip": {"num_bytes": 69966436, "checksum": "b7790d7ddc2f30ab1cebc094c166180e08e2520414c2f16003b615d302de67dd"}, "https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/somali/eval.zip": {"num_bytes": 7048257, "checksum": "9546a6010f36fcc8d75bedde57bd382ee89b7dea8df1cd8ff7f58dd9bc179ebe"}}, "download_size": 77014693, "post_processing_size": null, "dataset_size": 211213165, "size_in_bytes": 288227858}, "swahili": {"description": "Corpus used for training AfriBERTa models\n", "citation": "@inproceedings{ogueji-etal-2021-small,\n title = \"Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages\",\n author = \"Ogueji, Kelechi and\n Zhu, Yuxin and\n Lin, Jimmy\",\n booktitle = \"Proceedings of the 1st Workshop on Multilingual Representation Learning\",\n month = nov,\n year = \"2021\",\n address = \"Punta Cana, Dominican Republic\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/2021.mrl-1.11\",\n pages = \"116--126\",\n}\n", "homepage": "https://github.com/keleog/afriberta", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "afriberta_corpus", "config_name": "swahili", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 214063393, "num_examples": 1442912, "dataset_name": "afriberta_corpus"}, "test": {"name": "test", "num_bytes": 14699183, "num_examples": 100000, "dataset_name": "afriberta_corpus"}}, "download_checksums": {"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/swahili/train.zip": {"num_bytes": 78842932, "checksum": "0193b437552158b38b8affac9ed69bc19028f22d5511aef8802e96e6a94fcb64"}, "https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/swahili/eval.zip": {"num_bytes": 5465772, "checksum": "79fa9383ef2ee9b132a13978501423c2706b392c6a28d6562fc0439c8b8709f3"}}, "download_size": 84308704, "post_processing_size": null, "dataset_size": 228762576, "size_in_bytes": 313071280}, "tigrinya": {"description": "Corpus used for training AfriBERTa models\n", "citation": "@inproceedings{ogueji-etal-2021-small,\n title = \"Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages\",\n author = \"Ogueji, Kelechi and\n Zhu, Yuxin and\n Lin, Jimmy\",\n booktitle = \"Proceedings of the 1st Workshop on Multilingual Representation Learning\",\n month = nov,\n year = \"2021\",\n address = \"Punta Cana, Dominican Republic\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/2021.mrl-1.11\",\n pages = \"116--126\",\n}\n", "homepage": "https://github.com/keleog/afriberta", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "afriberta_corpus", "config_name": "tigrinya", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 28302512, "num_examples": 12076, "dataset_name": "afriberta_corpus"}, "test": {"name": "test", "num_bytes": 12114852, "num_examples": 5000, "dataset_name": "afriberta_corpus"}}, "download_checksums": {"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/tigrinya/train.zip": {"num_bytes": 6384012, "checksum": "d1713c6505a173db7bf7bf94eada1cf6446ea331f5d6bdd38d0d4d79386e878e"}, "https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/tigrinya/eval.zip": {"num_bytes": 2738799, "checksum": "e7e87a7152e35515469b79eb4ca2577968cea2f41b26384d638baed9bfe15e4d"}}, "download_size": 9122811, "post_processing_size": null, "dataset_size": 40417364, "size_in_bytes": 49540175}, "yoruba": {"description": "Corpus used for training AfriBERTa models\n", "citation": "@inproceedings{ogueji-etal-2021-small,\n title = \"Small Data? No Problem! Exploring the Viability of Pretrained Multilingual Language Models for Low-resourced Languages\",\n author = \"Ogueji, Kelechi and\n Zhu, Yuxin and\n Lin, Jimmy\",\n booktitle = \"Proceedings of the 1st Workshop on Multilingual Representation Learning\",\n month = nov,\n year = \"2021\",\n address = \"Punta Cana, Dominican Republic\",\n publisher = \"Association for Computational Linguistics\",\n url = \"https://aclanthology.org/2021.mrl-1.11\",\n pages = \"116--126\",\n}\n", "homepage": "https://github.com/keleog/afriberta", "license": "", "features": {"id": {"dtype": "string", "id": null, "_type": "Value"}, "text": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "afriberta_corpus", "config_name": "yoruba", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 30176961, "num_examples": 149148, "dataset_name": "afriberta_corpus"}, "test": {"name": "test", "num_bytes": 3019189, "num_examples": 15000, "dataset_name": "afriberta_corpus"}}, "download_checksums": {"https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/yoruba/train.zip": {"num_bytes": 10263359, "checksum": "701bf32e3b32ca3f6bccfa98f05d11de2fe0a23f9683d633e3e748591edf93ad"}, "https://huggingface.co/datasets/castorini/afriberta-corpus/resolve/main/yoruba/eval.zip": {"num_bytes": 1038458, "checksum": "e7d6bd2e34bb8d1af50ffdce7eb9912f4067dd8c01c4f348b7278e01697794a0"}}, "download_size": 11301817, "post_processing_size": null, "dataset_size": 33196150, "size_in_bytes": 44497967}}
|
|
|
|
gahuza/{eval.zip → afriberta-corpus-test.parquet}
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:046f510b97b5e2e2c39e3a2fc54d76346e259b7027027e493aff54b28fb8b586
|
3 |
+
size 1903244
|
amharic/train.zip → gahuza/afriberta-corpus-train.parquet
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2006b5cc3b14dc764ab37c2b1397f806e3311518778defec73d36af05100f677
|
3 |
+
size 16669795
|
gahuza/train.zip → hausa/afriberta-corpus-test.parquet
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bd43fb0b2f50c07431841491d28926280bd5c0e9fa1e6e09b0a0bd54753cef35
|
3 |
+
size 9144893
|
hausa/afriberta-corpus-train.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:445f932dabeca771ebbf54fc01182b1f6f325afecff68f618cad415c794c252d
|
3 |
+
size 117375246
|
hausa/eval.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:eced39cebce201f93cc0f36fe61437b0e1d11dd67628f17c5c19a057fb45b59a
|
3 |
-
size 4946084
|
|
|
|
|
|
|
|
hausa/train.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:7c6fa68b7714fe990c94d27b7825282905d2cb3baaddf5d4217b41eae0ffc605
|
3 |
-
size 63388604
|
|
|
|
|
|
|
|
igbo/afriberta-corpus-test.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e853483e630394acfca94d7a586adb046136d70e965eecabfa6d93c0d6866807
|
3 |
+
size 2592725
|
igbo/afriberta-corpus-train.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3d5aceaa4359aed50b8f9c7acbc0c9aad4c40afad99dba674d98438684ac4935
|
3 |
+
size 29091740
|
igbo/eval.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:4a43a72c19f136f3893ad7cde1f8588ae559358028295dec1f0267d561d503ba
|
3 |
-
size 1398211
|
|
|
|
|
|
|
|
igbo/train.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:0b42e54d602fdc47bbe736151222cf4e648340abebc49e0ab8143ef181a591de
|
3 |
-
size 15695174
|
|
|
|
|
|
|
|
pidgin/afriberta-corpus-test.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:86a334c3e56252958268aab0f20148e30e43407bc44944bc622f9770dc3d071c
|
3 |
+
size 2260451
|
pidgin/afriberta-corpus-train.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f8be8874928dc6cfc75486ef7b288b3c04875c67d0dfbdb0ea663fb7a1a1e415
|
3 |
+
size 24342861
|
pidgin/eval.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:495119efdf722d10aa69a404f0b2b05d8379318531fc0d018acfd0d9aef9b467
|
3 |
-
size 1220740
|
|
|
|
|
|
|
|
pidgin/train.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:6d52b73d4034f8cd88822b64b8bf3592fae54d89d58c13774b334b38c64034d8
|
3 |
-
size 13121362
|
|
|
|
|
|
|
|
somali/afriberta-corpus-test.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0faa78f48c9899c0850802f73fe1bbc7f7ff06ed805f274dd8a66bed37ca2880
|
3 |
+
size 12948602
|
somali/afriberta-corpus-train.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0f29c944b57a733ad527df3c53daec35d6c5d6145b641207af494fea575bea14
|
3 |
+
size 128581163
|
somali/eval.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:9546a6010f36fcc8d75bedde57bd382ee89b7dea8df1cd8ff7f58dd9bc179ebe
|
3 |
-
size 7048257
|
|
|
|
|
|
|
|
somali/train.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:b7790d7ddc2f30ab1cebc094c166180e08e2520414c2f16003b615d302de67dd
|
3 |
-
size 69966436
|
|
|
|
|
|
|
|
swahili/afriberta-corpus-test.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c5ba9b04ccfe3dfdd02849d056be95c37d5757025ad32c74584ad0f714c4d58b
|
3 |
+
size 10120734
|
swahili/afriberta-corpus-train.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:983266a953326607bcd551585d5cb7954d0c5894c550cdc930752e553f5a1d6c
|
3 |
+
size 145959473
|
swahili/eval.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:79fa9383ef2ee9b132a13978501423c2706b392c6a28d6562fc0439c8b8709f3
|
3 |
-
size 5465772
|
|
|
|
|
|
|
|
swahili/train.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:0193b437552158b38b8affac9ed69bc19028f22d5511aef8802e96e6a94fcb64
|
3 |
-
size 78842932
|
|
|
|
|
|
|
|
tigrinya/afriberta-corpus-test.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f5ab355e4dfec4944854f12d0404cd1130d91b89ffa696bb092916b77ba04954
|
3 |
+
size 5163589
|
tigrinya/afriberta-corpus-train.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1b37ec566ca6dd46dd8527fd424553195c2576c783404345da5285ae2e785c95
|
3 |
+
size 12015163
|
tigrinya/eval.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:e7e87a7152e35515469b79eb4ca2577968cea2f41b26384d638baed9bfe15e4d
|
3 |
-
size 2738799
|
|
|
|
|
|
|
|
tigrinya/train.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:d1713c6505a173db7bf7bf94eada1cf6446ea331f5d6bdd38d0d4d79386e878e
|
3 |
-
size 6384012
|
|
|
|
|
|
|
|
yoruba/afriberta-corpus-test.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6a48713c160a947791a8ee4b1e7360d41552d77e2afed1b8114921fd373605c7
|
3 |
+
size 1866580
|
yoruba/afriberta-corpus-train.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e13526d76c66935213f71774d57e391622c8d29a1cf10c9f25313c770cb4fada
|
3 |
+
size 18523363
|
yoruba/eval.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:e7d6bd2e34bb8d1af50ffdce7eb9912f4067dd8c01c4f348b7278e01697794a0
|
3 |
-
size 1038458
|
|
|
|
|
|
|
|
yoruba/train.zip
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:701bf32e3b32ca3f6bccfa98f05d11de2fe0a23f9683d633e3e748591edf93ad
|
3 |
-
size 10263359
|
|
|
|
|
|
|
|