ir_datasets: Mr. TyDiA multi-lingual benchmark benchmark suite constructed from the TyDi QA Benchmark. Relevance labels are sparsely assigned based on shallow human annotation.
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }Complete Arabic dataset, including all train, dev, and test queries and qrels.
Language: ar
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Language: ar
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 17K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 2106586,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 16595
},
"qrels": {
"count": 16749,
"fields": {
"relevance": {
"counts_by_value": {
"1": 16749
}
}
}
}
}
Development set for Arabic
Language: ar
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar/dev")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar/dev queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ar
Language: ar
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar/dev")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar/dev docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 3.1K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar/dev")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar/dev qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 2106586,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 3115
},
"qrels": {
"count": 3115,
"fields": {
"relevance": {
"counts_by_value": {
"1": 3115
}
}
}
}
}
Test set for Arabic
Language: ar
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar/test")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar/test queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ar
Language: ar
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar/test")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar/test docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 1.3K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar/test")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar/test qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 2106586,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 1081
},
"qrels": {
"count": 1257,
"fields": {
"relevance": {
"counts_by_value": {
"1": 1257
}
}
}
}
}
Train set for Arabic
Language: ar
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar/train")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar/train queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ar
Language: ar
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar/train")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar/train docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 12K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ar/train")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ar/train qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 2106586,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 12377
},
"qrels": {
"count": 12377,
"fields": {
"relevance": {
"counts_by_value": {
"1": 12377
}
}
}
}
}
Complete Bengali dataset, including all train, dev, and test queries and qrels.
Language: bn
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Language: bn
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 2.3K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 304059,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 2264
},
"qrels": {
"count": 2292,
"fields": {
"relevance": {
"counts_by_value": {
"1": 2292
}
}
}
}
}
Development set for Bengali
Language: bn
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn/dev")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn/dev queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/bn
Language: bn
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn/dev")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn/dev docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 443 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn/dev")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn/dev qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 304059,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 440
},
"qrels": {
"count": 443,
"fields": {
"relevance": {
"counts_by_value": {
"1": 443
}
}
}
}
}
Test set for Bengali
Language: bn
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn/test")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn/test queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/bn
Language: bn
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn/test")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn/test docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 130 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn/test")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn/test qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 304059,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 111
},
"qrels": {
"count": 130,
"fields": {
"relevance": {
"counts_by_value": {
"1": 130
}
}
}
}
}
Train set for Bengali
Language: bn
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn/train")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn/train queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/bn
Language: bn
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn/train")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn/train docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 1.7K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/bn/train")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/bn/train qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 304059,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 1713
},
"qrels": {
"count": 1719,
"fields": {
"relevance": {
"counts_by_value": {
"1": 1719
}
}
}
}
}
Complete English dataset, including all train, dev, and test queries and qrels.
Language: en
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en queries
[query_id] [text]
...
You can find more details about the CLI here.
import pyterrier as pt
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en')
index_ref = pt.IndexRef.of('./indices/mr-tydi_en') # assumes you have already built an index
pipeline = pt.BatchRetrieve(index_ref, wmodel='BM25')
# (optionally other pipeline components)
pipeline(dataset.get_topics())
You can find more details about PyTerrier retrieval here.
Language: en
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en docs
[doc_id] [text]
...
You can find more details about the CLI here.
import pyterrier as pt
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en')
# Index mr-tydi/en
indexer = pt.IterDictIndexer('./indices/mr-tydi_en')
index_ref = indexer.index(dataset.get_corpus_iter(), fields=['text'])
You can find more details about PyTerrier indexing here.
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 5.4K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
import pyterrier as pt
from pyterrier.measures import *
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en')
index_ref = pt.IndexRef.of('./indices/mr-tydi_en') # assumes you have already built an index
pipeline = pt.BatchRetrieve(index_ref, wmodel='BM25')
# (optionally other pipeline components)
pt.Experiment(
[pipeline],
dataset.get_topics(),
dataset.get_qrels(),
[MAP, nDCG@20]
)
You can find more details about PyTerrier experiments here.
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 32907100,
"fields": {
"doc_id": {
"max_len": 13,
"common_prefix": ""
}
}
},
"queries": {
"count": 5194
},
"qrels": {
"count": 5360,
"fields": {
"relevance": {
"counts_by_value": {
"1": 5360
}
}
}
}
}
Development set for English
Language: en
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en/dev")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en/dev queries
[query_id] [text]
...
You can find more details about the CLI here.
import pyterrier as pt
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en/dev')
index_ref = pt.IndexRef.of('./indices/mr-tydi_en') # assumes you have already built an index
pipeline = pt.BatchRetrieve(index_ref, wmodel='BM25')
# (optionally other pipeline components)
pipeline(dataset.get_topics())
You can find more details about PyTerrier retrieval here.
Inherits docs from mr-tydi/en
Language: en
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en/dev")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en/dev docs
[doc_id] [text]
...
You can find more details about the CLI here.
import pyterrier as pt
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en/dev')
# Index mr-tydi/en
indexer = pt.IterDictIndexer('./indices/mr-tydi_en')
index_ref = indexer.index(dataset.get_corpus_iter(), fields=['text'])
You can find more details about PyTerrier indexing here.
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 878 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en/dev")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en/dev qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
import pyterrier as pt
from pyterrier.measures import *
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en/dev')
index_ref = pt.IndexRef.of('./indices/mr-tydi_en') # assumes you have already built an index
pipeline = pt.BatchRetrieve(index_ref, wmodel='BM25')
# (optionally other pipeline components)
pt.Experiment(
[pipeline],
dataset.get_topics(),
dataset.get_qrels(),
[MAP, nDCG@20]
)
You can find more details about PyTerrier experiments here.
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 32907100,
"fields": {
"doc_id": {
"max_len": 13,
"common_prefix": ""
}
}
},
"queries": {
"count": 878
},
"qrels": {
"count": 878,
"fields": {
"relevance": {
"counts_by_value": {
"1": 878
}
}
}
}
}
Test set for English
Language: en
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en/test")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en/test queries
[query_id] [text]
...
You can find more details about the CLI here.
import pyterrier as pt
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en/test')
index_ref = pt.IndexRef.of('./indices/mr-tydi_en') # assumes you have already built an index
pipeline = pt.BatchRetrieve(index_ref, wmodel='BM25')
# (optionally other pipeline components)
pipeline(dataset.get_topics())
You can find more details about PyTerrier retrieval here.
Inherits docs from mr-tydi/en
Language: en
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en/test")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en/test docs
[doc_id] [text]
...
You can find more details about the CLI here.
import pyterrier as pt
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en/test')
# Index mr-tydi/en
indexer = pt.IterDictIndexer('./indices/mr-tydi_en')
index_ref = indexer.index(dataset.get_corpus_iter(), fields=['text'])
You can find more details about PyTerrier indexing here.
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 935 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en/test")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en/test qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
import pyterrier as pt
from pyterrier.measures import *
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en/test')
index_ref = pt.IndexRef.of('./indices/mr-tydi_en') # assumes you have already built an index
pipeline = pt.BatchRetrieve(index_ref, wmodel='BM25')
# (optionally other pipeline components)
pt.Experiment(
[pipeline],
dataset.get_topics(),
dataset.get_qrels(),
[MAP, nDCG@20]
)
You can find more details about PyTerrier experiments here.
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 32907100,
"fields": {
"doc_id": {
"max_len": 13,
"common_prefix": ""
}
}
},
"queries": {
"count": 744
},
"qrels": {
"count": 935,
"fields": {
"relevance": {
"counts_by_value": {
"1": 935
}
}
}
}
}
Train set for English
Language: en
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en/train")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en/train queries
[query_id] [text]
...
You can find more details about the CLI here.
import pyterrier as pt
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en/train')
index_ref = pt.IndexRef.of('./indices/mr-tydi_en') # assumes you have already built an index
pipeline = pt.BatchRetrieve(index_ref, wmodel='BM25')
# (optionally other pipeline components)
pipeline(dataset.get_topics())
You can find more details about PyTerrier retrieval here.
Inherits docs from mr-tydi/en
Language: en
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en/train")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en/train docs
[doc_id] [text]
...
You can find more details about the CLI here.
import pyterrier as pt
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en/train')
# Index mr-tydi/en
indexer = pt.IterDictIndexer('./indices/mr-tydi_en')
index_ref = indexer.index(dataset.get_corpus_iter(), fields=['text'])
You can find more details about PyTerrier indexing here.
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 3.5K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/en/train")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/en/train qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
import pyterrier as pt
from pyterrier.measures import *
pt.init()
dataset = pt.get_dataset('irds:mr-tydi/en/train')
index_ref = pt.IndexRef.of('./indices/mr-tydi_en') # assumes you have already built an index
pipeline = pt.BatchRetrieve(index_ref, wmodel='BM25')
# (optionally other pipeline components)
pt.Experiment(
[pipeline],
dataset.get_topics(),
dataset.get_qrels(),
[MAP, nDCG@20]
)
You can find more details about PyTerrier experiments here.
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 32907100,
"fields": {
"doc_id": {
"max_len": 13,
"common_prefix": ""
}
}
},
"queries": {
"count": 3547
},
"qrels": {
"count": 3547,
"fields": {
"relevance": {
"counts_by_value": {
"1": 3547
}
}
}
}
}
Complete Finnish dataset, including all train, dev, and test queries and qrels.
Language: fi
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Language: fi
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 9.8K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1908757,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 9572
},
"qrels": {
"count": 9750,
"fields": {
"relevance": {
"counts_by_value": {
"1": 9750
}
}
}
}
}
Development set for Finnish
Language: fi
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi/dev")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi/dev queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/fi
Language: fi
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi/dev")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi/dev docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 1.7K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi/dev")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi/dev qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1908757,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 1738
},
"qrels": {
"count": 1738,
"fields": {
"relevance": {
"counts_by_value": {
"1": 1738
}
}
}
}
}
Test set for Finnish
Language: fi
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi/test")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi/test queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/fi
Language: fi
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi/test")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi/test docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 1.5K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi/test")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi/test qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1908757,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 1254
},
"qrels": {
"count": 1451,
"fields": {
"relevance": {
"counts_by_value": {
"1": 1451
}
}
}
}
}
Train set for Finnish
Language: fi
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi/train")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi/train queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/fi
Language: fi
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi/train")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi/train docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 6.6K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/fi/train")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/fi/train qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1908757,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 6561
},
"qrels": {
"count": 6561,
"fields": {
"relevance": {
"counts_by_value": {
"1": 6561
}
}
}
}
}
Complete Indonesian dataset, including all train, dev, and test queries and qrels.
Language: id
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Language: id
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 7.1K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1469399,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 6977
},
"qrels": {
"count": 7087,
"fields": {
"relevance": {
"counts_by_value": {
"1": 7087
}
}
}
}
}
Development set for Indonesian
Language: id
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id/dev")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id/dev queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/id
Language: id
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id/dev")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id/dev docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 1.2K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id/dev")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id/dev qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1469399,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 1224
},
"qrels": {
"count": 1224,
"fields": {
"relevance": {
"counts_by_value": {
"1": 1224
}
}
}
}
}
Test set for Indonesian
Language: id
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id/test")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id/test queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/id
Language: id
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id/test")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id/test docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 961 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id/test")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id/test qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1469399,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 829
},
"qrels": {
"count": 961,
"fields": {
"relevance": {
"counts_by_value": {
"1": 961
}
}
}
}
}
Train set for Indonesian
Language: id
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id/train")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id/train queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/id
Language: id
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id/train")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id/train docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 4.9K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/id/train")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/id/train qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1469399,
"fields": {
"doc_id": {
"max_len": 11,
"common_prefix": ""
}
}
},
"queries": {
"count": 4902
},
"qrels": {
"count": 4902,
"fields": {
"relevance": {
"counts_by_value": {
"1": 4902
}
}
}
}
}
Complete Japanese dataset, including all train, dev, and test queries and qrels.
Language: ja
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Language: ja
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 5.5K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 7000027,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 5353
},
"qrels": {
"count": 5548,
"fields": {
"relevance": {
"counts_by_value": {
"1": 5548
}
}
}
}
}
Development set for Japanese
Language: ja
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja/dev")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja/dev queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ja
Language: ja
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja/dev")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja/dev docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 928 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja/dev")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja/dev qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 7000027,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 928
},
"qrels": {
"count": 928,
"fields": {
"relevance": {
"counts_by_value": {
"1": 928
}
}
}
}
}
Test set for Japanese
Language: ja
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja/test")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja/test queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ja
Language: ja
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja/test")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja/test docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 923 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja/test")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja/test qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 7000027,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 720
},
"qrels": {
"count": 923,
"fields": {
"relevance": {
"counts_by_value": {
"1": 923
}
}
}
}
}
Train set for Japanese
Language: ja
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja/train")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja/train queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ja
Language: ja
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja/train")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja/train docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 3.7K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ja/train")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ja/train qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 7000027,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 3697
},
"qrels": {
"count": 3697,
"fields": {
"relevance": {
"counts_by_value": {
"1": 3697
}
}
}
}
}
Complete Korean dataset, including all train, dev, and test queries and qrels.
Language: ko
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Language: ko
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 2.1K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1496126,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 2019
},
"qrels": {
"count": 2116,
"fields": {
"relevance": {
"counts_by_value": {
"1": 2116
}
}
}
}
}
Development set for Korean
Language: ko
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko/dev")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko/dev queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ko
Language: ko
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko/dev")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko/dev docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 307 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko/dev")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko/dev qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1496126,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 303
},
"qrels": {
"count": 307,
"fields": {
"relevance": {
"counts_by_value": {
"1": 307
}
}
}
}
}
Test set for Korean
Language: ko
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko/test")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko/test queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ko
Language: ko
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko/test")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko/test docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 492 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko/test")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko/test qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1496126,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 421
},
"qrels": {
"count": 492,
"fields": {
"relevance": {
"counts_by_value": {
"1": 492
}
}
}
}
}
Train set for Korean
Language: ko
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko/train")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko/train queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ko
Language: ko
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko/train")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko/train docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 1.3K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ko/train")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ko/train qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 1496126,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 1295
},
"qrels": {
"count": 1317,
"fields": {
"relevance": {
"counts_by_value": {
"1": 1317
}
}
}
}
}
Complete Russian dataset, including all train, dev, and test queries and qrels.
Language: ru
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Language: ru
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 7.9K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 9597504,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 7763
},
"qrels": {
"count": 7909,
"fields": {
"relevance": {
"counts_by_value": {
"1": 7909
}
}
}
}
}
Development set for Russian
Language: ru
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru/dev")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru/dev queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ru
Language: ru
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru/dev")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru/dev docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 1.4K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru/dev")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru/dev qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 9597504,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 1375
},
"qrels": {
"count": 1375,
"fields": {
"relevance": {
"counts_by_value": {
"1": 1375
}
}
}
}
}
Test set for Russian
Language: ru
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru/test")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru/test queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ru
Language: ru
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru/test")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru/test docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 1.2K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru/test")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru/test qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 9597504,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 995
},
"qrels": {
"count": 1168,
"fields": {
"relevance": {
"counts_by_value": {
"1": 1168
}
}
}
}
}
Train set for Russian
Language: ru
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru/train")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru/train queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/ru
Language: ru
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru/train")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru/train docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 5.4K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/ru/train")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/ru/train qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 9597504,
"fields": {
"doc_id": {
"max_len": 12,
"common_prefix": ""
}
}
},
"queries": {
"count": 5366
},
"qrels": {
"count": 5366,
"fields": {
"relevance": {
"counts_by_value": {
"1": 5366
}
}
}
}
}
Complete Swahili dataset, including all train, dev, and test queries and qrels.
Language: sw
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Language: sw
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 3.8K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 136689,
"fields": {
"doc_id": {
"max_len": 9,
"common_prefix": ""
}
}
},
"queries": {
"count": 3271
},
"qrels": {
"count": 3767,
"fields": {
"relevance": {
"counts_by_value": {
"1": 3767
}
}
}
}
}
Development set for Swahili
Language: sw
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw/dev")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw/dev queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/sw
Language: sw
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw/dev")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw/dev docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 623 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw/dev")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw/dev qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 136689,
"fields": {
"doc_id": {
"max_len": 9,
"common_prefix": ""
}
}
},
"queries": {
"count": 526
},
"qrels": {
"count": 623,
"fields": {
"relevance": {
"counts_by_value": {
"1": 623
}
}
}
}
}
Test set for Swahili
Language: sw
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw/test")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw/test queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/sw
Language: sw
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw/test")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw/test docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 743 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw/test")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw/test qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 136689,
"fields": {
"doc_id": {
"max_len": 9,
"common_prefix": ""
}
}
},
"queries": {
"count": 670
},
"qrels": {
"count": 743,
"fields": {
"relevance": {
"counts_by_value": {
"1": 743
}
}
}
}
}
Train set for Swahili
Language: sw
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw/train")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw/train queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/sw
Language: sw
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw/train")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw/train docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 2.4K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/sw/train")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/sw/train qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 136689,
"fields": {
"doc_id": {
"max_len": 9,
"common_prefix": ""
}
}
},
"queries": {
"count": 2072
},
"qrels": {
"count": 2401,
"fields": {
"relevance": {
"counts_by_value": {
"1": 2401
}
}
}
}
}
Complete Telugu dataset, including all train, dev, and test queries and qrels.
Language: te
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Language: te
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 5.5K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 548224,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 5517
},
"qrels": {
"count": 5540,
"fields": {
"relevance": {
"counts_by_value": {
"1": 5540
}
}
}
}
}
Development set for Telugu
Language: te
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te/dev")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te/dev queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/te
Language: te
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te/dev")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te/dev docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 983 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te/dev")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te/dev qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 548224,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 983
},
"qrels": {
"count": 983,
"fields": {
"relevance": {
"counts_by_value": {
"1": 983
}
}
}
}
}
Test set for Telugu
Language: te
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te/test")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te/test queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/te
Language: te
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te/test")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te/test docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 677 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te/test")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te/test qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 548224,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 646
},
"qrels": {
"count": 677,
"fields": {
"relevance": {
"counts_by_value": {
"1": 677
}
}
}
}
}
Train set for Telugu
Language: te
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te/train")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te/train queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/te
Language: te
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te/train")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te/train docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 3.9K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/te/train")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/te/train qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 548224,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 3880
},
"qrels": {
"count": 3880,
"fields": {
"relevance": {
"counts_by_value": {
"1": 3880
}
}
}
}
}
Complete Thai dataset, including all train, dev, and test queries and qrels.
Language: th
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Language: th
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 5.5K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 568855,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 5322
},
"qrels": {
"count": 5545,
"fields": {
"relevance": {
"counts_by_value": {
"1": 5545
}
}
}
}
}
Development set for Thai
Language: th
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th/dev")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th/dev queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/th
Language: th
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th/dev")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th/dev docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 817 | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th/dev")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th/dev qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 568855,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 807
},
"qrels": {
"count": 817,
"fields": {
"relevance": {
"counts_by_value": {
"1": 817
}
}
}
}
}
Test set for Thai
Language: th
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th/test")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th/test queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/th
Language: th
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th/test")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th/test docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 1.4K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th/test")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th/test qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 568855,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 1190
},
"qrels": {
"count": 1368,
"fields": {
"relevance": {
"counts_by_value": {
"1": 1368
}
}
}
}
}
Train set for Thai
Language: th
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th/train")
for query in dataset.queries_iter():
query # namedtuple<query_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th/train queries
[query_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Inherits docs from mr-tydi/th
Language: th
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th/train")
for doc in dataset.docs_iter():
doc # namedtuple<doc_id, text>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th/train docs
[doc_id] [text]
...
You can find more details about the CLI here.
No example available for PyTerrier
Relevance levels
| Rel. | Definition | Count | % |
|---|---|---|---|
| 1 | Passage identified within Wikipedia article from top Google search results | 3.4K | 100.0% |
Examples:
import ir_datasets
dataset = ir_datasets.load("mr-tydi/th/train")
for qrel in dataset.qrels_iter():
qrel # namedtuple<query_id, doc_id, relevance, iteration>
You can find more details about the Python API here.
ir_datasets export mr-tydi/th/train qrels --format tsv
[query_id] [doc_id] [relevance] [iteration]
...
You can find more details about the CLI here.
No example available for PyTerrier
Bibtex:
@article{Zhang2021MrTyDi, title={{Mr. TyDi}: A Multi-lingual Benchmark for Dense Retrieval}, author={Xinyu Zhang and Xueguang Ma and Peng Shi and Jimmy Lin}, year={2021}, journal={arXiv:2108.08787}, } @article{Clark2020TyDiQa, title={{TyDi QA}: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages}, author={Jonathan H. Clark and Eunsol Choi and Michael Collins and Dan Garrette and Tom Kwiatkowski and Vitaly Nikolaev and Jennimaria Palomaki}, year={2020}, journal={Transactions of the Association for Computational Linguistics} }{
"docs": {
"count": 568855,
"fields": {
"doc_id": {
"max_len": 10,
"common_prefix": ""
}
}
},
"queries": {
"count": 3319
},
"qrels": {
"count": 3360,
"fields": {
"relevance": {
"counts_by_value": {
"1": 3360
}
}
}
}
}