holylovenia
commited on
Commit
•
4880225
1
Parent(s):
8199987
Upload roots_vi_ted.py with huggingface_hub
Browse files- roots_vi_ted.py +128 -0
roots_vi_ted.py
ADDED
@@ -0,0 +1,128 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from pathlib import Path
|
2 |
+
from typing import Dict, List, Tuple
|
3 |
+
|
4 |
+
import datasets
|
5 |
+
import pandas as pd
|
6 |
+
|
7 |
+
from seacrowd.utils import schemas
|
8 |
+
from seacrowd.utils.configs import SEACrowdConfig
|
9 |
+
from seacrowd.utils.constants import Licenses, Tasks
|
10 |
+
|
11 |
+
_CITATION = """
|
12 |
+
@inproceedings{DBLP:conf/nips/LaurenconSWAMSW22,
|
13 |
+
author={Hugo Laurençon and Lucile Saulnier and Thomas Wang and Christopher Akiki and Albert Villanova del Moral and
|
14 |
+
Teven Le Scao and Leandro von Werra and Chenghao Mou and Eduardo González Ponferrada and Huu Nguyen and Jörg Frohberg
|
15 |
+
and Mario Sasko and Quentin Lhoest and Angelina McMillan-Major and Gérard Dupont and Stella Biderman and Anna Rogers
|
16 |
+
and Loubna Ben Allal and Francesco De Toni and Giada Pistilli and Olivier Nguyen and Somaieh Nikpoor and Maraim Masoud
|
17 |
+
and Pierre Colombo and Javier de la Rosa and Paulo Villegas and Tristan Thrush and Shayne Longpre and Sebastian Nagel
|
18 |
+
and Leon Weber and Manuel Muñoz and Jian Zhu and Daniel van Strien and Zaid Alyafeai and Khalid Almubarak and Minh
|
19 |
+
Chien Vu and Itziar Gonzalez-Dios and Aitor Soroa and Kyle Lo and Manan Dey and Pedro Ortiz Suarez and Aaron Gokaslan
|
20 |
+
and Shamik Bose and David Ifeoluwa Adelani and Long Phan and Hieu Tran and Ian Yu and Suhas Pai and Jenny Chim and
|
21 |
+
Violette Lepercq and Suzana Ilic and Margaret Mitchell and Alexandra Sasha Luccioni and Yacine Jernite},
|
22 |
+
title={The BigScience ROOTS Corpus: A 1.6TB Composite Multilingual Dataset},
|
23 |
+
year={2022},
|
24 |
+
cdate={1640995200000},
|
25 |
+
url={http://papers.nips.cc/paper_files/paper/2022/hash/ce9e92e3de2372a4b93353eb7f3dc0bd-Abstract-Datasets_and_Benchmarks.html},
|
26 |
+
booktitle={NeurIPS},
|
27 |
+
}
|
28 |
+
"""
|
29 |
+
|
30 |
+
_DATASETNAME = "roots_vi_ted"
|
31 |
+
|
32 |
+
_DESCRIPTION = """
|
33 |
+
ROOTS_vi_ted is a subset of Vietnamese in ted_talks_iwslt datasets. ted_talks_iwslt is a collection of the original Ted
|
34 |
+
talks and their translated version. The translations are available in more than 109+ languages, though the distribution
|
35 |
+
is not uniform. Before using this dataloader, please accept the acknowledgement at
|
36 |
+
https://huggingface.co/datasets/bigscience-data/roots_vi_ted_talks_iwslt and use huggingface-cli login for authentication.
|
37 |
+
"""
|
38 |
+
|
39 |
+
_HOMEPAGE = "https://huggingface.co/datasets/bigscience-data/roots_vi_ted_talks_iwslt"
|
40 |
+
|
41 |
+
_LANGUAGES = ["vie"]
|
42 |
+
|
43 |
+
_LICENSE = Licenses.CC_BY_NC_ND_4_0.value
|
44 |
+
|
45 |
+
_LOCAL = False
|
46 |
+
|
47 |
+
_URLS = {_DATASETNAME: {"train": "https://huggingface.co/datasets/bigscience-data/roots_vi_ted_talks_iwslt/resolve/main/data/train-00000-of-00001.parquet?download=true"}}
|
48 |
+
|
49 |
+
_SUPPORTED_TASKS = [Tasks.SELF_SUPERVISED_PRETRAINING]
|
50 |
+
|
51 |
+
_SOURCE_VERSION = "1.0.0"
|
52 |
+
|
53 |
+
_SEACROWD_VERSION = "2024.06.20"
|
54 |
+
|
55 |
+
|
56 |
+
class RootsViTedDataset(datasets.GeneratorBasedBuilder):
|
57 |
+
"""RootsViTed is a subset of Vietnamese in ted_talks_iwslt datasets."""
|
58 |
+
|
59 |
+
SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
|
60 |
+
SEACROWD_VERSION = datasets.Version(_SEACROWD_VERSION)
|
61 |
+
|
62 |
+
BUILDER_CONFIGS = [
|
63 |
+
SEACrowdConfig(
|
64 |
+
name="roots_vi_ted_source",
|
65 |
+
version=SOURCE_VERSION,
|
66 |
+
description="roots_vi_ted source schema",
|
67 |
+
schema="source",
|
68 |
+
subset_id="roots_vi_ted",
|
69 |
+
),
|
70 |
+
SEACrowdConfig(
|
71 |
+
name="roots_vi_ted_seacrowd_ssp",
|
72 |
+
version=SEACROWD_VERSION,
|
73 |
+
description="roots_vi_ted SEACrowd schema",
|
74 |
+
schema="seacrowd_ssp",
|
75 |
+
subset_id="roots_vi_ted",
|
76 |
+
),
|
77 |
+
]
|
78 |
+
|
79 |
+
DEFAULT_CONFIG_NAME = "roots_vi_ted_source"
|
80 |
+
|
81 |
+
def _info(self) -> datasets.DatasetInfo:
|
82 |
+
if self.config.schema == "source":
|
83 |
+
features = datasets.Features(
|
84 |
+
{
|
85 |
+
"text": datasets.Value("string"),
|
86 |
+
"meta": datasets.Value("string"),
|
87 |
+
}
|
88 |
+
)
|
89 |
+
|
90 |
+
elif self.config.schema == "seacrowd_ssp":
|
91 |
+
features = schemas.self_supervised_pretraining.features
|
92 |
+
|
93 |
+
return datasets.DatasetInfo(
|
94 |
+
description=_DESCRIPTION,
|
95 |
+
features=features,
|
96 |
+
homepage=_HOMEPAGE,
|
97 |
+
license=_LICENSE,
|
98 |
+
citation=_CITATION,
|
99 |
+
)
|
100 |
+
|
101 |
+
def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
|
102 |
+
"""Returns SplitGenerators."""
|
103 |
+
urls = _URLS[_DATASETNAME]
|
104 |
+
data_dir = dl_manager.download_and_extract(urls)
|
105 |
+
|
106 |
+
return [
|
107 |
+
datasets.SplitGenerator(
|
108 |
+
name=datasets.Split.TRAIN,
|
109 |
+
gen_kwargs={"filepath": data_dir, "split": "train"},
|
110 |
+
),
|
111 |
+
]
|
112 |
+
|
113 |
+
def _generate_examples(self, filepath: Path, split: str) -> Tuple[int, Dict]:
|
114 |
+
if self.config.schema == "source":
|
115 |
+
df = pd.read_parquet(filepath[split])
|
116 |
+
for i, row in df.iterrows():
|
117 |
+
yield i, {
|
118 |
+
"text": row["text"],
|
119 |
+
"meta": row["meta"],
|
120 |
+
}
|
121 |
+
|
122 |
+
elif self.config.schema == "seacrowd_ssp":
|
123 |
+
df = pd.read_parquet(filepath[split])
|
124 |
+
for i, row in df.iterrows():
|
125 |
+
yield i, {
|
126 |
+
"id": str(i),
|
127 |
+
"text": row["text"],
|
128 |
+
}
|