File size: 4,505 Bytes
f65642d 5602cf0 f65642d |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 |
# coding=utf-8
# Lint as: python3
"""XWinograd"""
import json
import pandas as pd
import datasets
logger = datasets.logging.get_logger(__name__)
_CITATION = """\
@misc{tikhonov2021heads,
title={It's All in the Heads: Using Attention Heads as a Baseline for Cross-Lingual Transfer in Commonsense Reasoning},
author={Alexey Tikhonov and Max Ryabinin},
year={2021},
eprint={2106.12066},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
"""
_DESCRIPTION = """\
A multilingual collection of Winograd Schemas in six languages \
that can be used for evaluation of cross-lingual commonsense reasoning capabilities.
"""
#_URL = "https://github.com/yandex-research/crosslingual_winograd/blob/main/dataset.tsv"
#_URL = "https://huggingface.co/datasets/muennighoff/xwinograd/resolve/main/data/xwinograd.tsv"
_URL = "https://huggingface.co/datasets/Muennighoff/xwinograd/blob/main/data/xwinograd.tsv"
import json
import random
def winogrande_format(row):
array = row["pronoun"]
position_idx = json.loads(array)[1][0]
# Turn unicode into proper chinese characters
sent = str(u"{}".format(row["sent"]))
start_idx = 0
for i, tok in enumerate(json.loads(row["toks"])):
tok = str(u"{}".format(tok))
cur_start_idx = sent.find(tok)
if i == position_idx:
break
sent = sent[cur_start_idx + len(tok):]
start_idx += cur_start_idx + len(tok)
# +1 to give room for an optional space
row["sentence"] = row["sent"][:start_idx] + row["sent"][start_idx:start_idx+len(tok)+1].replace(tok, "_") + row["sent"][start_idx+len(tok)+1:]
sol = json.loads(row["solution"])
cor_answer_idx = random.choice([1, 2])
incor_answer_idx = 2 if cor_answer_idx == 1 else 1
cor_answer = str(u"{}".format(sol[0][0])) if sol[0][-1] == True else str(u"{}".format(sol[1][0]))
incor_answer = str(u"{}".format(sol[0][0])) if sol[0][-1] == False else str(u"{}".format(sol[1][0]))
row[f"option{cor_answer_idx}"] = cor_answer
row[f"option{incor_answer_idx}"] = incor_answer
row["answer"] = cor_answer_idx
return row
class XWinograd(datasets.GeneratorBasedBuilder):
"""XWinograd"""
VERSION = datasets.Version("1.0.0")
BUILDER_CONFIGS = [
datasets.BuilderConfig(
name="en",
version=VERSION,
description="X",
),
datasets.BuilderConfig(
name="fr",
version=VERSION,
description="X",
),
datasets.BuilderConfig(
name="jp",
version=VERSION,
description="X",
),
datasets.BuilderConfig(
name="pt",
version=VERSION,
description="X",
),
datasets.BuilderConfig(
name="ru",
version=VERSION,
description="X",
),
datasets.BuilderConfig(
name="zh",
version=VERSION,
description="X",
),
]
DEFAULT_CONFIG_NAME = "en"
def _info(self):
return datasets.DatasetInfo(
description=_DESCRIPTION,
features=datasets.Features(
{
"sentence": datasets.Value("string"),
"option1": datasets.Value("string"),
"option2": datasets.Value("string"),
"answer": datasets.Value("string")
}
),
supervised_keys=None,
citation=_CITATION,
)
def _split_generators(self, dl_manager):
downloaded_files = dl_manager.download_and_extract(_URL)
return [
datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs={"filepath": downloaded_files}),
]
def _generate_examples(self, filepath):
"""This function returns the examples in the raw (text) form."""
logger.info("generating examples from = %s", filepath)
ds = pd.read_csv(
filepath, sep='\t', header=None,
names=["lang", "type", "original", "sent", "toks", "pronoun", "solution"]
)
if self.config.name:
ds = ds[ds["lang"] == self.config.name]
ds = ds.apply(winogrande_format, axis=1)
for idx, row in ds.iterrows():
yield idx, {
"sentence": row["sentence"],
"option1": row["option1"],
"option2": row["option2"],
"answer": row["answer"],
}
|