1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148 | """All German dataset configurations used in EuroEval."""
from ..data_models import DatasetConfig
from ..languages import DE
from ..tasks import COMMON_SENSE, EUROPEAN_VALUES, KNOW, LA, MCRC, NER, RC, SENT, SUMM
### Official datasets ###
SB10K_CONFIG = DatasetConfig(
name="sb10k",
pretty_name="the truncated version of the German sentiment classification "
"dataset SB10k",
huggingface_id="EuroEval/sb10k-mini",
task=SENT,
languages=[DE],
)
SCALA_DE_CONFIG = DatasetConfig(
name="scala-de",
pretty_name="the German part of the linguistic acceptability dataset ScaLA",
huggingface_id="EuroEval/scala-de",
task=LA,
languages=[DE],
)
GERMEVAL_CONFIG = DatasetConfig(
name="germeval",
pretty_name="the truncated version of the German named entity recognition "
"dataset GermEval",
huggingface_id="EuroEval/germeval-mini",
task=NER,
languages=[DE],
)
GERMANQUAD_CONFIG = DatasetConfig(
name="germanquad",
pretty_name="the truncated version of the German reading comprehension dataset "
"GermanQuAD",
huggingface_id="EuroEval/germanquad-mini",
task=RC,
languages=[DE],
)
MLSUM_DE_CONFIG = DatasetConfig(
name="mlsum-de",
pretty_name="the truncated version of the German summarisation dataset MLSum-de",
huggingface_id="EuroEval/mlsum-mini",
task=SUMM,
languages=[DE],
)
MMLU_DE_CONFIG = DatasetConfig(
name="mmlu-de",
pretty_name="the truncated version of the German knowledge dataset MMLU-de, "
"translated from the English MMLU dataset",
huggingface_id="EuroEval/mmlu-de-mini",
task=KNOW,
languages=[DE],
)
HELLASWAG_DE_CONFIG = DatasetConfig(
name="hellaswag-de",
pretty_name="the truncated version of the German common-sense reasoning dataset "
"HellaSwag-de, translated from the English HellaSwag dataset",
huggingface_id="EuroEval/hellaswag-de-mini",
task=COMMON_SENSE,
languages=[DE],
)
EUROPEAN_VALUES_DE_CONFIG = DatasetConfig(
name="european-values-de",
pretty_name="the German version of the European values evaluation dataset",
huggingface_id="EuroEval/european-values-de",
task=EUROPEAN_VALUES,
languages=[DE],
splits=["test"],
bootstrap_samples=False,
_instruction_prompt="{text}",
)
### Unofficial datasets ###
ARC_DE_CONFIG = DatasetConfig(
name="arc-de",
pretty_name="the truncated version of the German knowledge dataset ARC-de, "
"translated from the English ARC dataset",
huggingface_id="EuroEval/arc-de-mini",
task=KNOW,
languages=[DE],
unofficial=True,
)
BELEBELE_DE_CONFIG = DatasetConfig(
name="belebele-de",
pretty_name="the German multiple choice reading comprehension dataset BeleBele-de, "
"translated from the English BeleBele dataset",
huggingface_id="EuroEval/belebele-de-mini",
task=MCRC,
languages=[DE],
unofficial=True,
)
MULTI_WIKI_QA_DE_CONFIG = DatasetConfig(
name="multi-wiki-qa-de",
pretty_name="the truncated version of the German part of the reading "
"comprehension dataset MultiWikiQA",
huggingface_id="EuroEval/multi-wiki-qa-de-mini",
task=RC,
languages=[DE],
unofficial=True,
)
GOLDENSWAG_DE_CONFIG = DatasetConfig(
name="goldenswag-de",
pretty_name="the truncated version of the German common-sense reasoning "
"dataset GoldenSwag-de, translated from the English GoldenSwag dataset",
huggingface_id="EuroEval/goldenswag-de-mini",
task=COMMON_SENSE,
languages=[DE],
unofficial=True,
)
EUROPEAN_VALUES_SITUATIONAL_DE_CONFIG = DatasetConfig(
name="european-values-situational-de",
pretty_name="the German version of the European values evaluation dataset, where "
"the questions are phrased in a situational way",
huggingface_id="EuroEval/european-values-situational-de",
task=EUROPEAN_VALUES,
languages=[DE],
splits=["test"],
bootstrap_samples=False,
_instruction_prompt="{text}",
unofficial=True,
)
EUROPEAN_VALUES_COMPLETIONS_DE_CONFIG = DatasetConfig(
name="european-values-completions-de",
pretty_name="the German version of the European values evaluation dataset, where "
"the questions are phrased as sentence completions",
huggingface_id="EuroEval/european-values-completions-de",
task=EUROPEAN_VALUES,
languages=[DE],
splits=["test"],
bootstrap_samples=False,
_instruction_prompt="{text}",
unofficial=True,
)
|