diff --git "a/dataset_infos.json" "b/dataset_infos.json"
new file mode 100644--- /dev/null
+++ "b/dataset_infos.json"
@@ -0,0 +1 @@
+{"afr-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "afr-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "afr-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "afr-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "afr-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "afr-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "afr-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "afr-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "afr-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "afr-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "afr-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "afr-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ain-fin": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ain-fin", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-ber": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-ber", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-ber_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-ber_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-ell": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-ell", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-heb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-heb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ara-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ara-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "arq-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "arq-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "avk-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "avk-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "avk-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "avk-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "awa-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "awa-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "aze-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "aze-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "aze-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "aze-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "aze-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "aze-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "aze_Latn-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "aze_Latn-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bel-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bel-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ben-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ben-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ber-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ber-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ber-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ber-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ber-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ber-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ber-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ber-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ber-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ber-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ber_Latn-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ber_Latn-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ber_Latn-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ber_Latn-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ber_Latn-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ber_Latn-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ber_Latn-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ber_Latn-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bre-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bre-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bre-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bre-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bua-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bua-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bua_Cyrl-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bua_Cyrl-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-bul": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-bul", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "bul-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "bul-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cat-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cat-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cat-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cat-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cat-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cat-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cat-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cat-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cat-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cat-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cat-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cat-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cat-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cat-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cat-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cat-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cat-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cat-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cbk-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cbk-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ceb-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ceb-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ceb-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ceb-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-hun": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-hun", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-slv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-slv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ces-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ces-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cha-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cha-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "chm-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "chm-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "chv-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "chv-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "chv-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "chv-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "chv-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "chv-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cmn_Hans-wuu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cmn_Hans-wuu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cor-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cor-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cor-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cor-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cor-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cor-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cor-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cor-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cor-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cor-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cor-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cor-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cor-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cor-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "crh-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "crh-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "cym-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "cym-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-dan": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-dan", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-fin": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-fin", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-nob": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-nob", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dan-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dan-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-deu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-deu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-dsb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-dsb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-ell": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-ell", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-est": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-est", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-eus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-eus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-fas": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-fas", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-fin": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-fin", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-frr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-frr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-gos": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-gos", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-hbs": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-hbs", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-heb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-heb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-hrv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-hrv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-hrx": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-hrx", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-hsb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-hsb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-hun": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-hun", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-ido": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-ido", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-ile": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-ile", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-ina": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-ina", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-ind": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-ind", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-isl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-isl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-jbo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-jbo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-jpn_Hani": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-jpn_Hani", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-jpn_Kana": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-jpn_Kana", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-kab": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-kab", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-kor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-kor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-kor_Hang": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-kor_Hang", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-kur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-kur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-kur_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-kur_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-lad": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-lad", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-lfn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-lfn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-lfn_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-lfn_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-lit": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-lit", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-ltz": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-ltz", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-msa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-msa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-nds": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-nds", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-nob": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-nob", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-ron": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-ron", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-run": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-run", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-slv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-slv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-srp_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-srp_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-swg": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-swg", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-tat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-tat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-tgl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-tgl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-tlh": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-tlh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-toki": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-toki", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-vie": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-vie", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-vol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-vol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "deu-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "deu-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dsb-hsb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dsb-hsb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dsb-slv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dsb-slv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dtp-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dtp-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dtp-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dtp-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dtp-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dtp-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dtp-msa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dtp-msa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "dtp-zsm_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "dtp-zsm_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "egl-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "egl-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ell-ell": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ell-ell", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ell-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ell-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ell-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ell-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ell-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ell-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ell-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ell-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ell-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ell-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ell-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ell-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ell-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ell-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ell-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ell-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ell-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ell-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ell-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ell-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-bos_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-bos_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-eng": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-eng", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-est": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-est", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-eus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-eus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-fao": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-fao", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-fas": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-fas", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-fin": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-fin", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-fry": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-fry", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-gla": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-gla", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-gle": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-gle", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-glg": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-glg", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-gos": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-gos", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-got": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-got", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-grc": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-grc", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-gsw": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-gsw", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-hbs": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-hbs", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-heb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-heb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-hin": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-hin", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-hoc": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-hoc", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-hoc_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-hoc_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-hrv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-hrv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-hrx": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-hrx", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-hun": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-hun", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-hye": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-hye", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ido": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ido", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ido_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ido_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ile": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ile", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ilo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ilo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ina": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ina", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ind": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ind", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-isl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-isl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-jav": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-jav", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-jbo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-jbo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-jbo_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-jbo_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-jpn_Hani": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-jpn_Hani", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-jpn_Kana": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-jpn_Kana", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-kab": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-kab", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-kat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-kat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-kaz": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-kaz", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-kaz_Cyrl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-kaz_Cyrl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-kha": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-kha", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-khm": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-khm", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-kor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-kor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-kor_Hang": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-kor_Hang", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-kur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-kur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-kur_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-kur_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-kzj": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-kzj", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-lad": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-lad", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-lad_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-lad_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-lav": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-lav", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-lfn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-lfn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-lfn_Cyrl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-lfn_Cyrl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-lfn_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-lfn_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-lit": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-lit", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ltz": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ltz", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-mal": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-mal", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-mar": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-mar", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-mkd": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-mkd", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-mlt": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-mlt", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-mon": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-mon", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-mri": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-mri", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-msa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-msa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-mya": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-mya", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-nds": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-nds", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-nno": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-nno", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-nob": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-nob", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-nov": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-nov", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-nst": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-nst", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-oci": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-oci", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-orv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-orv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ota": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ota", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ota_Arab": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ota_Arab", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ota_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ota_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-pam": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-pam", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-pes": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-pes", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-pms": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-pms", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-prg": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-prg", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-que": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-que", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-rom": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-rom", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ron": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ron", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-run": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-run", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-slv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-slv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-sqi": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-sqi", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-srp_Cyrl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-srp_Cyrl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-srp_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-srp_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-swa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-swa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-tam": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-tam", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-tat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-tat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-tel": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-tel", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-tgl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-tgl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-tha": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-tha", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-tlh": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-tlh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-toki": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-toki", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-tuk": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-tuk", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-tuk_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-tuk_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-tzl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-tzl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-tzl_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-tzl_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-uig": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-uig", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-uig_Arab": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-uig_Arab", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-urd": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-urd", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-uzb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-uzb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-uzb_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-uzb_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-vie": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-vie", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-vol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-vol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-war": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-war", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-xal": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-xal", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-yue_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-yue_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-yue_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-yue_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-zsm_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-zsm_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eng-zza": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eng-zza", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-epo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-epo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-fas": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-fas", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-fin": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-fin", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-glg": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-glg", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-hbs": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-hbs", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-heb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-heb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-hrv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-hrv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-hun": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-hun", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-ido": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-ido", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-ile": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-ile", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-ile_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-ile_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-ina": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-ina", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-isl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-isl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-jbo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-jbo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-jpn_Hani": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-jpn_Hani", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-lad": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-lad", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-lad_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-lad_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-lfn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-lfn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-lfn_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-lfn_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-lit": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-lit", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-nds": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-nds", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-nob": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-nob", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-oci": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-oci", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-ron": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-ron", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-slv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-slv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-srp_Cyrl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-srp_Cyrl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-srp_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-srp_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-tgl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-tgl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-tlh": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-tlh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-toki": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-toki", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-vie": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-vie", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-vol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-vol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "epo-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "epo-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "est-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "est-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eus-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eus-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eus-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eus-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "eus-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "eus-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fas-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fas-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-fin": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-fin", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-fkv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-fkv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-heb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-heb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-hun": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-hun", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-jpn_Hani": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-jpn_Hani", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-jpn_Kana": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-jpn_Kana", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-kor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-kor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-kor_Hang": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-kor_Hang", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-kur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-kur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-nno": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-nno", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-nob": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-nob", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fin-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fin-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-fra": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-fra", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-gcf": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-gcf", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-hbs": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-hbs", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-heb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-heb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-hrv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-hrv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-hun": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-hun", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-ido": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-ido", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-ile": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-ile", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-ina": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-ina", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-ind": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-ind", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-jbo": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-jbo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-jpn_Hani": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-jpn_Hani", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-kab": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-kab", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-kor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-kor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-kor_Hang": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-kor_Hang", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-lfn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-lfn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-lfn_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-lfn_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-msa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-msa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-nds": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-nds", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-nob": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-nob", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-oci": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-oci", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-pcd": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-pcd", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-ron": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-ron", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-run": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-run", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-slv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-slv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-tat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-tat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-tgl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-tgl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-tlh": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-tlh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-tlh_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-tlh_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-toki": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-toki", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-toki_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-toki_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-uig": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-uig", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-uig_Arab": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-uig_Arab", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-vie": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-vie", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-wuu": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-wuu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fra-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fra-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "fry-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "fry-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "gcf-gcf": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "gcf-gcf", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "gla-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "gla-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "glg-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "glg-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "glg-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "glg-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "gos-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "gos-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "grn-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "grn-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "grn-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "grn-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hbs-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hbs-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hbs-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hbs-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hbs-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hbs-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hbs-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hbs-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hbs-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hbs-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hbs-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hbs-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hbs-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hbs-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hbs-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hbs-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-heb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-heb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-hun": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-hun", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-ina": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-ina", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-lad": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-lad", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-lfn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-lfn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-lfn_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-lfn_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "heb-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "heb-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hin-urd": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hin-urd", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hin-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hin-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hrv-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hrv-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hrv-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hrv-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hrv-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hrv-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hrv-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hrv-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hsb-slv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hsb-slv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-hun": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-hun", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-jpn_Hani": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-jpn_Hani", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-kor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-kor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-kor_Hang": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-kor_Hang", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hun-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hun-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "hye-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "hye-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ido-ina": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ido-ina", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ido-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ido-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ido-lfn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ido-lfn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ido-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ido-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ido-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ido-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ido_Latn-lfn_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ido_Latn-lfn_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina-lad": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina-lad", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina-lfn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina-lfn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina-tlh": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina-tlh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina_Latn-lad_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina_Latn-lad_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina_Latn-lfn_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina_Latn-lfn_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ina_Latn-tlh_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ina_Latn-tlh_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ind-zsm_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ind-zsm_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "isl-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "isl-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "isl-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "isl-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "isl-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "isl-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "isl-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "isl-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-ind": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-ind", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-jpn_Hani": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-jpn_Hani", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-lit": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-lit", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-msa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-msa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-nds": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-nds", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-pms": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-pms", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-ron": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-ron", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-toki": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-toki", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-vie": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-vie", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ita-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ita-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jbo-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jbo-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jbo-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jbo-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jbo-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jbo-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jbo-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jbo-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jbo-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jbo-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jbo_Latn-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jbo_Latn-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jbo_Latn-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jbo_Latn-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jbo_Latn-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jbo_Latn-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-jpn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-jpn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-kor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-kor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-lit": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-lit", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-mar": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-mar", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-msa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-msa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-nds": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-nds", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-tlh": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-tlh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-toki": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-toki", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-vie": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-vie", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hani-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hani-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hani-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hani-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hani-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hani-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hani-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hani-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hani-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hani-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hani-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hani-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-ind": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-ind", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-jpn_Hira": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-jpn_Hira", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-kor_Hang": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-kor_Hang", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-lit": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-lit", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-mar": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-mar", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-nds": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-nds", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-nob": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-nob", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-tlh_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-tlh_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Hira-vie": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Hira-vie", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Kana-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Kana-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "jpn_Kana-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "jpn_Kana-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kab-kab": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kab-kab", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kab-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kab-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kab-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kab-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kat-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kat-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kaz-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kaz-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kaz_Cyrl-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kaz_Cyrl-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "khm-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "khm-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kor-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kor-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kor-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kor-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kor-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kor-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kor_Hang-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kor_Hang-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kor_Hang-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kor_Hang-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kor_Hang-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kor_Hang-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kzj-msa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kzj-msa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "kzj_Latn-zsm_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "kzj_Latn-zsm_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lad-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lad-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lad-lfn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lad-lfn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lad-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lad-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lad-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lad-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lad_Latn-lfn_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lad_Latn-lfn_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lad_Latn-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lad_Latn-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lad_Latn-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lad_Latn-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat-lat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat-lat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat-lfn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat-lfn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat-tlh": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat-tlh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat_Latn-lfn_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat_Latn-lfn_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lat_Latn-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lat_Latn-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lav-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lav-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lfn-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lfn-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lfn-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lfn-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lfn-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lfn-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lfn-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lfn-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lfn_Cyrl-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lfn_Cyrl-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lfn_Latn-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lfn_Latn-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lfn_Latn-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lfn_Latn-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lit-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lit-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lit-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lit-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lit-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lit-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "lit-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "lit-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ltz-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ltz-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "mkd-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "mkd-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "msa-msa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "msa-msa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "msa-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "msa-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "msa-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "msa-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nds-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nds-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nds-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nds-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nds-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nds-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nds-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nds-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-nld": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-nld", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-ron": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-ron", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-toki": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-toki", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nld-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nld-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nno-nob": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nno-nob", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nob-nno": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nob-nno", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nob-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nob-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nob-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nob-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nob-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nob-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nor-nor": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nor-nor", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nor-pol": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nor-pol", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nor-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nor-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nor-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nor-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nor-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nor-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nor-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nor-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nor-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nor-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "nor-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "nor-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "orv-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "orv-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ota-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ota-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "pol-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "pol-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "pol-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "pol-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "pol-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "pol-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "pol-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "pol-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "pol-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "pol-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "pol-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "pol-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "pol-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "pol-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "pol-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "pol-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "pol-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "pol-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-por": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-por", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-ron": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-ron", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-tgl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-tgl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-toki": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-toki", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "por-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "por-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ron-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ron-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ron-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ron-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ron-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ron-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "run-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "run-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "run-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "run-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-sah": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-sah", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-slv": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-slv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-tat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-tat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-tlh": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-tlh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-toki": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-toki", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-toki_Latn": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-toki_Latn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-uig": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-uig", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-uig_Arab": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-uig_Arab", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-vie": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-vie", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-xal": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-xal", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-yue_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-yue_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "rus-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "rus-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "slv-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "slv-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "slv-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "slv-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "slv-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "slv-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-spa": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-spa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-tat": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-tat", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-tgl": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-tgl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-tlh": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-tlh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-toki": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-toki", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-vie": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-vie", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "spa-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "spa-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "srp_Cyrl-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "srp_Cyrl-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "srp_Cyrl-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "srp_Cyrl-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "srp_Latn-ita": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "srp_Latn-ita", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "srp_Latn-nob": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "srp_Latn-nob", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "srp_Latn-rus": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "srp_Latn-rus", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "srp_Latn-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "srp_Latn-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "swe-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "swe-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "swe-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "swe-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "swe-swe": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "swe-swe", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "swe-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "swe-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "swe-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "swe-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tat-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tat-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tat-vie": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tat-vie", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tlh-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tlh-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tlh-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tlh-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tlh_Latn-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tlh_Latn-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tlh_Latn-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tlh_Latn-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tlh_Latn-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tlh_Latn-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tur-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tur-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tur-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tur-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tur-tur": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tur-tur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tur-uig": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tur-uig", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tur-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tur-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tur-uzb": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tur-uzb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "tur-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "tur-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "uig-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "uig-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "uig_Arab-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "uig_Arab-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "uig_Arab-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "uig_Arab-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ukr-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ukr-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ukr-cmn_Hant": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ukr-cmn_Hant", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ukr-ukr": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ukr-ukr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "ukr-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "ukr-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "vie-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "vie-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "vie-vie": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "vie-vie", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "vie-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "vie-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "wuu-cmn_Hans": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "wuu-cmn_Hans", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "yid-yid": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "yid-yid", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "zho-zho": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "zho-zho", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}, "zsm_Latn-ind": {"description": "The Tatoeba Translation Challenge is a multilingual data set of\nmachine translation benchmarks derived from user-contributed\ntranslations collected by [Tatoeba.org](https://tatoeba.org/) and\nprovided as parallel corpus from [OPUS](https://opus.nlpl.eu/). This\ndataset includes test and development data sorted by language pair. It\nincludes test sets for hundreds of language pairs and is continuously\nupdated. Please, check the version number tag to refer to the release\nthat your are using.\n", "citation": "@inproceedings{tiedemann-2020-tatoeba,\n    title = \"The {T}atoeba {T}ranslation {C}hallenge {--} {R}ealistic Data Sets for Low Resource and Multilingual {MT}\",\n    author = {Tiedemann, J{\"o}rg},\n    booktitle = \"Proceedings of the Fifth Conference on Machine Translation\",\n    month = nov,\n    year = \"2020\",\n    publisher = \"Association for Computational Linguistics\",\n    url = \"https://aclanthology.org/2020.wmt-1.139\",\n    pages = \"1174--1182\",\n}\n", "homepage": "https://github.com/Helsinki-NLP/Tatoeba-Challenge/", "license": "", "features": {"sourceLang": {"dtype": "string", "id": null, "_type": "Value"}, "targetlang": {"dtype": "string", "id": null, "_type": "Value"}, "sourceString": {"dtype": "string", "id": null, "_type": "Value"}, "targetString": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "builder_name": "tatoeba_mt", "config_name": "zsm_Latn-ind", "version": "0.0.0", "splits": {}, "download_checksums": {}, "download_size": 0, "post_processing_size": null, "dataset_size": 0, "size_in_bytes": 0}}
\ No newline at end of file