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 | """All Portuguese dataset configurations used in EuroEval."""
from ..data_models import DatasetConfig
from ..languages import PT
from ..tasks import COMMON_SENSE, EUROPEAN_VALUES, KNOW, LA, MCRC, NER, RC, SENT, SUMM
### Official datasets ###
SST2_PT_CONFIG = DatasetConfig(
name="sst2-pt",
pretty_name="the truncated version of the Portuguese sentiment classification "
"dataset SST2-pt, translated from the English SST2 dataset",
huggingface_id="EuroEval/sst2-pt-mini",
task=SENT,
languages=[PT],
_labels=["positive", "negative"],
)
SCALA_PT = DatasetConfig(
name="scala-pt",
pretty_name="the Portuguese part of the linguistic acceptability dataset ScaLA",
huggingface_id="EuroEval/scala-pt",
task=LA,
languages=[PT],
)
HAREM_CONFIG = DatasetConfig(
name="harem",
pretty_name="the Portuguese named entity recognition dataset HAREM",
huggingface_id="EuroEval/harem",
task=NER,
languages=[PT],
)
MULTI_WIKI_QA_PT_CONFIG = DatasetConfig(
name="multi-wiki-qa-pt",
pretty_name="the truncated version of the Portuguese part of the reading "
"comprehension dataset MultiWikiQA",
huggingface_id="EuroEval/multi-wiki-qa-pt-pt-mini",
task=RC,
languages=[PT],
)
PUBLICO_CONFIG = DatasetConfig(
name="publico",
pretty_name="the truncated version of the Portuguese summarisation dataset Público",
huggingface_id="EuroEval/publico-mini",
task=SUMM,
languages=[PT],
)
MMLU_PT_CONFIG = DatasetConfig(
name="mmlu-pt",
pretty_name="the truncated version of the Portuguese knowledge dataset MMLU-pt, "
"translated from the English MMLU dataset",
huggingface_id="EuroEval/mmlu-pt-mini",
task=KNOW,
languages=[PT],
)
GOLDENSWAG_PT_CONFIG = DatasetConfig(
name="goldenswag-pt",
pretty_name="the truncated version of the Portuguese common-sense reasoning "
"dataset GoldenSwag-pt, translated from the English GoldenSwag dataset",
huggingface_id="EuroEval/goldenswag-pt-mini",
task=COMMON_SENSE,
languages=[PT],
)
EUROPEAN_VALUES_PT_CONFIG = DatasetConfig(
name="european-values-pt",
pretty_name="the Portuguese version of the European values evaluation dataset",
huggingface_id="EuroEval/european-values-pt",
task=EUROPEAN_VALUES,
languages=[PT],
splits=["test"],
bootstrap_samples=False,
_instruction_prompt="{text}",
)
### Unofficial datasets ###
BOOLQ_PT_CONFIG = DatasetConfig(
name="boolq-pt",
pretty_name="the Portuguese multiple choice reading comprehension dataset "
"BoolQ-pt, translated from the English BoolQ dataset",
huggingface_id="EuroEval/boolq-pt",
task=MCRC,
languages=[PT],
unofficial=True,
)
WINOGRANDE_PT_CONFIG = DatasetConfig(
name="winogrande-pt",
pretty_name="the Portuguese common-sense reasoning dataset Winogrande-pt, "
"translated from the English Winogrande dataset",
huggingface_id="EuroEval/winogrande-pt",
task=COMMON_SENSE,
languages=[PT],
_labels=["a", "b"],
unofficial=True,
)
EUROPEAN_VALUES_SITUATIONAL_PT_CONFIG = DatasetConfig(
name="european-values-situational-pt",
pretty_name="the Portuguese version of the European values evaluation dataset, "
"where the questions are phrased in a situational way",
huggingface_id="EuroEval/european-values-situational-pt",
task=EUROPEAN_VALUES,
languages=[PT],
splits=["test"],
bootstrap_samples=False,
_instruction_prompt="{text}",
unofficial=True,
)
EUROPEAN_VALUES_COMPLETIONS_PT_CONFIG = DatasetConfig(
name="european-values-completions-pt",
pretty_name="the Portuguese version of the European values evaluation dataset, "
"where the questions are phrased as sentence completions",
huggingface_id="EuroEval/european-values-completions-pt",
task=EUROPEAN_VALUES,
languages=[PT],
splits=["test"],
bootstrap_samples=False,
_instruction_prompt="{text}",
unofficial=True,
)
|