def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
"wikidata_id": nlp.Value("string"),
"text": nlp.Value("string"),
"version_id": nlp.Value("string"),
}),
supervised_keys=None,
homepage=_URL,
citation=_CITATION,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
_QUESTION: nlp.Value("string"),
_QUERY: nlp.Value("string"),
}),
supervised_keys=(_QUESTION, _QUERY),
homepage=
"https://github.com/google-research/google-research/tree/master/cfq",
citation=_CITATION,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
_DOCUMENT: nlp.Value("string"),
_SUMMARY: nlp.Value("string"),
"section_names": nlp.Value("string"),
}),
supervised_keys=None,
homepage="https://github.com/armancohan/long-summarization",
citation=_CITATION,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
"id": nlp.Value("string"),
"text": nlp.Value("string")
}),
supervised_keys=None,
homepage=_URL,
citation=_CITATION,
license=_LICENSE,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
"title": nlp.Value("string"),
"text": nlp.Value("string"),
}),
# No default supervised_keys.
supervised_keys=None,
homepage="https://dumps.wikimedia.org",
citation=_CITATION,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
_DOCUMENT: nlp.Value("string"),
_SUMMARY: nlp.Value("string"),
}),
supervised_keys=(_DOCUMENT, _SUMMARY),
homepage=
"https://github.com/EdinburghNLP/XSum/tree/master/XSum-Dataset",
citation=_CITATION,
)
def _info(self):
# TODO(scifact): Specifies the nlp.DatasetInfo object
if self.config.name == 'corpus':
features = {
"doc_id":
nlp.Value('int32'), # The document's S2ORC ID.
"title":
nlp.Value('string'), # The title.
"abstract":
nlp.features.Sequence({
'sentence': nlp.Value('string')
}), # The abstract, written as a list of sentences.
"structured":
nlp.Value(
'bool'
) # Indicator for whether this is a structured abstract.
}
else:
features = {
"id":
nlp.Value('int32'), # An integer claim ID.
"claim":
nlp.Value('string'), # The text of the claim.
"evidence_doc_id":
nlp.Value('string'),
"evidence_label":
nlp.Value('string'), # Label for the rationale.
"evidence_sentences":
nlp.features.Sequence({
'sentence': nlp.Value('int32') # Rationale sentences.
}),
"cited_doc_ids":
nlp.features.Sequence({'doc_id': nlp.Value('int32')
}) # The claim's "cited documents".
}
return nlp.DatasetInfo(
# This is the description that will appear on the datasets page.
description=_DESCRIPTION,
# nlp.features.FeatureConnectors
features=nlp.Features(
features
# These are the features of your dataset like images, labels ...
),
# If there's a common (input, target) tuple from the features,
# specify them here. They'll be used if as_supervised=True in
# builder.as_dataset.
supervised_keys=None,
# Homepage of the dataset for documentation
homepage='https://scifact.apps.allenai.org/',
citation=_CITATION,
)
def _info(self):
return nlp.MetricInfo(
description=_DESCRIPTION,
citation=_CITATION,
inputs_description=_KWARGS_DESCRIPTION,
features=nlp.Features({
'predictions': nlp.Sequence(nlp.Value('string', id='token'), id='sequence'),
'references': nlp.Sequence(nlp.Sequence(nlp.Value('string', id='token'), id='sequence'), id='references'),
}),
codebase_urls=["https://github.com/tensorflow/nmt/blob/master/nmt/scripts/bleu.py"],
reference_urls=["https://en.wikipedia.org/wiki/BLEU",
"https://towardsdatascience.com/evaluating-text-output-in-nlp-bleu-at-your-own-risk-e8609665a213"]
)
def _info(self):
features = {text_feature: nlp.Value("string") for text_feature in six.iterkeys(self.config.text_features)}
if self.config.label_classes:
features["label"] = nlp.features.ClassLabel(names=self.config.label_classes)
else:
features["label"] = nlp.Value("float32")
features["idx"] = nlp.Value("int32")
return nlp.DatasetInfo(
description=_GLUE_DESCRIPTION,
features=nlp.Features(features),
homepage=self.config.url,
citation=self.config.citation + "\n" + _GLUE_CITATION,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
"source_text": nlp.Value("string"),
"target_text": nlp.Value("string")
}),
# No default supervised_keys (as we have to pass both question
# and context as input).
supervised_keys=None,
homepage="https://rajpurkar.github.io/SQuAD-explorer/",
citation=_CITATION,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
"wikidata_id": nlp.Value('string'),
"text": nlp.Value('string'),
"version_id": nlp.Value('string'),
}),
supervised_keys=None,
homepage=_URL,
citation=_CITATION,
redistribution_info={"license": _LICENSE},
)
def _info(self):
# Should return a nlp.DatasetInfo object
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
_ARTICLE: nlp.Value("string"),
_HIGHLIGHTS: nlp.Value("string"),
"id": nlp.Value("string"),
}),
supervised_keys=None,
homepage="https://github.com/abisee/cnn-dailymail",
citation=_CITATION,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
_REVIEW_SENTS:
nlp.Value("string"),
_SUMMARIES:
nlp.features.Sequence(nlp.Value("string"))
}),
supervised_keys=(_REVIEW_SENTS, _SUMMARIES),
homepage="http://kavita-ganesan.com/opinosis/",
citation=_CITATION,
)
def _info(self):
return nlp.MetricInfo(
description=_DESCRIPTION,
citation=_CITATION,
homepage="https://github.com/chakki-works/seqeval",
inputs_description=_KWARGS_DESCRIPTION,
features=nlp.Features({
'predictions':
nlp.Sequence(nlp.Value('string', id='label'), id='sequence'),
'references':
nlp.Sequence(nlp.Value('string', id='label'), id='sequence'),
}),
codebase_urls=["https://github.com/chakki-works/seqeval"],
reference_urls=["https://github.com/chakki-works/seqeval"])
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features(
{
"id": nlp.Value("string"),
"tokens": nlp.Sequence(nlp.Value("string")),
"labels": nlp.Sequence(nlp.Value("string")),
}
),
supervised_keys=None,
homepage="http://noisy-text.github.io/2017/emerging-rare-entities.html",
citation=_CITATION,
)
def _info(self):
# TODO(cosmos_qa): Specifies the nlp.DatasetInfo object
return nlp.DatasetInfo(
# This is the description that will appear on the datasets page.
description=_DESCRIPTION,
# nlp.features.FeatureConnectors
features=nlp.Features({
"id": nlp.Value("string"),
"context": nlp.Value("string"),
"question": nlp.Value("string"),
"answer0": nlp.Value("string"),
"answer1": nlp.Value("string"),
"answer2": nlp.Value("string"),
"answer3": nlp.Value("string"),
"label": nlp.Value("int32")
# These are the features of your dataset like images, labels ...
}),
# If there's a common (input, target) tuple from the features,
# specify them here. They'll be used if as_supervised=True in
# builder.as_dataset.
supervised_keys=None,
# Homepage of the dataset for documentation
homepage="https://wilburone.github.io/cosmos/",
citation=_CITATION,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
"source_text": nlp.Value("string"),
"target_text": nlp.Value("string"),
"task": nlp.Value("string"),
}),
# No default supervised_keys (as we have to pass both question
# and context as input).
supervised_keys=None,
homepage="http://jmcauley.ucsd.edu/data/amazon/qa/",
citation=_CITATION,
)
def _info(self):
return nlp.MetricInfo(
description=_DESCRIPTION,
citation=_CITATION,
inputs_description=_KWARGS_DESCRIPTION,
features=nlp.Features({
'predictions': nlp.Value('string', id='sequence'),
'references': nlp.Value('string', id='sequence'),
}),
codebase_urls=["https://github.com/ns-moosavi/coval"],
reference_urls=["https://github.com/ns-moosavi/coval",
"https://www.aclweb.org/anthology/P16-1060",
"http://www.conll.cemantix.org/2012/data.html"]
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
"chunk":
nlp.Value("string"),
"chunk_id":
nlp.Value("int32"),
"turn_start":
nlp.Value("int32"),
"turn_end":
nlp.Value("int32"),
"alignment_score":
nlp.Value("float32"),
"turn_num":
nlp.Value("int32"),
"turns":
nlp.features.Sequence({
"names": nlp.Value("string"),
"utterances": nlp.Value("string"),
}),
}),
homepage="https://github.com/RevanthRameshkumar/CRD3",
citation=_CITATION,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
"context": nlp.Value("string"),
"answer": nlp.Value("string"),
"question": nlp.Value("string")
}),
# No default supervised_keys (as we have to pass both question
# and context as input).
supervised_keys=None,
homepage="https://datasets.maluuba.com/NewsQA",
citation=_CITATION,
)
def _info(self):
# TODO(discofuse): Specifies the nlp.DatasetInfo object
return nlp.DatasetInfo(
# This is the description that will appear on the datasets page.
description=_DESCRIPTION,
# nlp.features.FeatureConnectors
features=nlp.Features(
{
"connective_string": nlp.Value("string"),
"discourse_type": nlp.Value("string"),
"coherent_second_sentence": nlp.Value("string"),
"has_coref_type_pronoun": nlp.Value("float32"),
"incoherent_first_sentence": nlp.Value("string"),
"incoherent_second_sentence": nlp.Value("string"),
"has_coref_type_nominal": nlp.Value("float32"),
"coherent_first_sentence": nlp.Value("string"),
# These are the features of your dataset like images, labels ...
}
),
# If there's a common (input, target) tuple from the features,
# specify them here. They'll be used if as_supervised=True in
# builder.as_dataset.
supervised_keys=None,
# Homepage of the dataset for documentation
homepage="https://github.com/google-research-datasets/discofuse",
citation=_CITATION,
)
def _info(self):
# TODO(empathetic_dialogues): Specifies the nlp.DatasetInfo object
return nlp.DatasetInfo(
# This is the description that will appear on the datasets page.
description=_DESCRIPTION,
# nlp.features.FeatureConnectors
features=nlp.Features({
"conv_id": nlp.Value("string"),
"utterance_idx": nlp.Value("int32"),
"context": nlp.Value("string"),
"prompt": nlp.Value("string"),
"speaker_idx": nlp.Value("int32"),
"utterance": nlp.Value("string"),
"selfeval": nlp.Value("string"),
"tags": nlp.Value("string")
# These are the features of your dataset like images, labels ...
}),
# If there's a common (input, target) tuple from the features,
# specify them here. They'll be used if as_supervised=True in
# builder.as_dataset.
supervised_keys=None,
# Homepage of the dataset for documentation
homepage="https://github.com/facebookresearch/EmpatheticDialogues",
citation=_CITATION,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
"premise": nlp.Value("string"),
"hypothesis": nlp.Value("string"),
"label": nlp.Value("string"),
}),
# No default supervised_keys (as we have to pass both premise
# and hypothesis as input).
supervised_keys=None,
homepage="https://www.nyu.edu/projects/bowman/multinli/",
citation=_CITATION,
)
def _info(self):
return nlp.MetricInfo(
description=_DESCRIPTION,
citation=_CITATION,
homepage="https://github.com/Tiiiger/bert_score",
inputs_description=_KWARGS_DESCRIPTION,
features=nlp.Features({
'predictions': nlp.Value('string', id='sequence'),
'references': nlp.Sequence(nlp.Value('string', id='sequence'), id='references'),
}),
codebase_urls=["https://github.com/Tiiiger/bert_score"],
reference_urls=["https://github.com/Tiiiger/bert_score",
"https://arxiv.org/abs/1904.09675"]
)
def _info(self):
# TODO(jeopardy): Specifies the nlp.DatasetInfo object
return nlp.DatasetInfo(
# This is the description that will appear on the datasets page.
description=_DESCRIPTION,
# nlp.features.FeatureConnectors
features=nlp.Features(
{
"category": nlp.Value("string"),
"air_date": nlp.Value("string"),
"question": nlp.Value("string"),
"value": nlp.Value("int32"),
"answer": nlp.Value("string"),
"round": nlp.Value("string"),
"category": nlp.Value("string"),
"show_number": nlp.Value("int32"),
# These are the features of your dataset like images, labels ...
}
),
# If there's a common (input, target) tuple from the features,
# specify them here. They'll be used if as_supervised=True in
# builder.as_dataset.
supervised_keys=None,
# Homepage of the dataset for documentation
homepage=_URL,
citation=_CITATION,
)
def _info(self):
info = nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
_TITLE: nlp.Value("string"),
_DOCUMENT: nlp.Value("string"),
_SUMMARY: nlp.Value("string"),
_CLEAN_DOCUMENT: nlp.Value("string"),
_CLEAN_SUMMARY: nlp.Value("string"),
}),
# supervised_keys=(_TITLE, _DOCUMENT, _SUMMARY),
homepage="https://github.com/airKlizz/MultiDocMultiLingualSum",
citation=_CITATION,
)
return info
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features({
_QUESTION: nlp.Value('string'),
_QUERY: nlp.Value('string'),
}),
supervised_keys={
"input": _QUESTION,
"output": _QUERY
},
homepage=
'https://github.com/google-research/google-research/tree/master/cfq',
citation=_CITATION,
)
def _info(self):
features = {
k: nlp.Value("string")
for k in [_DOCUMENT, _SUMMARY] + _ADDITIONAL_TEXT_FEATURES
}
features.update(
{k: nlp.Value("float32")
for k in _ADDITIONAL_FLOAT_FEATURES})
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features(features),
supervised_keys=(_DOCUMENT, _SUMMARY),
homepage="http://lil.nlp.cornell.edu/newsroom/",
citation=_CITATION,
)
def _info(self):
return nlp.DatasetInfo(
description=_DESCRIPTION + self.config.description,
features=nlp.Features({
"input_data": nlp.Value("string"),
"label": nlp.Value("int32"),
"label_level_1": nlp.Value("int32"),
"label_level_2": nlp.Value("int32"),
}),
# No default supervised_keys (as we have to pass both premise
# and hypothesis as input).
supervised_keys=None,
homepage="https://data.mendeley.com/datasets/9rw3vkcfy4/6",
citation=_CITATION,
)
def _info(self):
features = {
"text": nlp.Value("string"),
"url": nlp.Value("string"),
"content-type": nlp.Value("string"),
"content-length": nlp.Value("string"),
"timestamp": nlp.Value("string"),
}
return nlp.DatasetInfo(
description=_DESCRIPTION,
features=nlp.Features(features),
citation=_CITATION,
homepage=
"https://github.com/google-research/text-to-text-transfer-transformer#datasets",
)