opus-100 / dataset_infos.json
system's picture
system HF staff
Update files from the datasets library (from 1.2.0)
8f48d18
raw
history blame
No virus
192 kB
{"af-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["af", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "af", "output": "en"}, "builder_name": "opus100", "config_name": "af-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 135916, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 18726471, "num_examples": 275512, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 132777, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-af-en-v1.0.tar.gz": {"num_bytes": 7505036, "checksum": "d5b1ac19fdc41508fdfd946e1c51ae6777d181474f5e77b1fee88bf76d254d52"}}, "download_size": 7505036, "post_processing_size": null, "dataset_size": 18995164, "size_in_bytes": 26500200}, "am-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["am", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "am", "output": "en"}, "builder_name": "opus100", "config_name": "am-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 588029, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 21950644, "num_examples": 89027, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 566077, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-am-en-v1.0.tar.gz": {"num_bytes": 7004193, "checksum": "c68d7590fd4c708d6b80639389cf73bd1ebc3ca39bbaeaaaf7c6d48140c780c4"}}, "download_size": 7004193, "post_processing_size": null, "dataset_size": 23104750, "size_in_bytes": 30108943}, "an-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["an", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "an", "output": "en"}, "builder_name": "opus100", "config_name": "an-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 438332, "num_examples": 6961, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-an-en-v1.0.tar.gz": {"num_bytes": 96148, "checksum": "8083ade0f12af4abf1600f497f4327af45a4d074cf85134a82e07bd7dfdff876"}}, "download_size": 96148, "post_processing_size": null, "dataset_size": 438332, "size_in_bytes": 534480}, "ar-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["ar", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "ar", "output": "en"}, "builder_name": "opus100", "config_name": "ar-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 331648, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 152766484, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 2272106, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-ar-en-v1.0.tar.gz": {"num_bytes": 55286865, "checksum": "efc55eaab636e5c469b1ab3e2543bcfdc57f375ee6d57f248cfcfa589b46594b"}}, "download_size": 55286865, "post_processing_size": null, "dataset_size": 155370238, "size_in_bytes": 210657103}, "as-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["as", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "as", "output": "en"}, "builder_name": "opus100", "config_name": "as-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 261466, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 15634648, "num_examples": 138479, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 248139, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-as-en-v1.0.tar.gz": {"num_bytes": 4183517, "checksum": "adf38305e6256bb5e338b6c9d9dd08b5755acb637b255196dda1b78e23e130dd"}}, "download_size": 4183517, "post_processing_size": null, "dataset_size": 16144253, "size_in_bytes": 20327770}, "az-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["az", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "az", "output": "en"}, "builder_name": "opus100", "config_name": "az-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 393109, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 56431259, "num_examples": 262089, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 407109, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-az-en-v1.0.tar.gz": {"num_bytes": 18897341, "checksum": "a3a3ba50c98ffa8cfb453d0023be3da9f4e5c6cf03ce03ae3c402b1c8745c7ab"}}, "download_size": 18897341, "post_processing_size": null, "dataset_size": 57231477, "size_in_bytes": 76128818}, "be-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["be", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "be", "output": "en"}, "builder_name": "opus100", "config_name": "be-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 166858, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 5298500, "num_examples": 67312, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 175205, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-be-en-v1.0.tar.gz": {"num_bytes": 1906088, "checksum": "f5a047de898b222f413bea7316ef60fbad99a5f4ab11b3041e1ea0d9d4439dab"}}, "download_size": 1906088, "post_processing_size": null, "dataset_size": 5640563, "size_in_bytes": 7546651}, "bg-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["bg", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "bg", "output": "en"}, "builder_name": "opus100", "config_name": "bg-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 243751, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 108930347, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 234848, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-bg-en-v1.0.tar.gz": {"num_bytes": 36980744, "checksum": "ee019090235379a17e90422fe739667b3af913da1ee55e662db10db317d6ec5a"}}, "download_size": 36980744, "post_processing_size": null, "dataset_size": 109408946, "size_in_bytes": 146389690}, "bn-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["bn", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "bn", "output": "en"}, "builder_name": "opus100", "config_name": "bn-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 510101, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 249906846, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 498414, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-bn-en-v1.0.tar.gz": {"num_bytes": 72999655, "checksum": "03cecc49791aef3788ebe6bd36908462e3e198916d42ca1165c74c7443580845"}}, "download_size": 72999655, "post_processing_size": null, "dataset_size": 250915361, "size_in_bytes": 323915016}, "br-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["br", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "br", "output": "en"}, "builder_name": "opus100", "config_name": "br-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 127925, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 8539006, "num_examples": 153447, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 133772, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-br-en-v1.0.tar.gz": {"num_bytes": 3323458, "checksum": "1dde74dd2f935df7bb5db8909c77bd0e19cef97e60195c9e9c2154a8f270af08"}}, "download_size": 3323458, "post_processing_size": null, "dataset_size": 8800703, "size_in_bytes": 12124161}, "bs-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["bs", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "bs", "output": "en"}, "builder_name": "opus100", "config_name": "bs-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 168622, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 75082948, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 172481, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-bs-en-v1.0.tar.gz": {"num_bytes": 30746956, "checksum": "de8747c8b5dd34f5e19aa9ccf3734626cf7f7c0b82ef4920a33e3c722acb70e2"}}, "download_size": 30746956, "post_processing_size": null, "dataset_size": 75424051, "size_in_bytes": 106171007}, "ca-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["ca", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "ca", "output": "en"}, "builder_name": "opus100", "config_name": "ca-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 205666, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 88405510, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 212637, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-ca-en-v1.0.tar.gz": {"num_bytes": 36267794, "checksum": "40aa2a23d4fb5408b05f9a2f95544e7b7195f7e0aa405fac93263930a77ab19a"}}, "download_size": 36267794, "post_processing_size": null, "dataset_size": 88823813, "size_in_bytes": 125091607}, "cs-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["cs", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "cs", "output": "en"}, "builder_name": "opus100", "config_name": "cs-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 205274, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 91897719, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 219084, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-cs-en-v1.0.tar.gz": {"num_bytes": 39673827, "checksum": "e870a377d0eb7327ad9f5e4d979feb41ad0e0a331bc176fcdd7e45ecdda1ff0f"}}, "download_size": 39673827, "post_processing_size": null, "dataset_size": 92322077, "size_in_bytes": 131995904}, "cy-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["cy", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "cy", "output": "en"}, "builder_name": "opus100", "config_name": "cy-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 124289, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 17244980, "num_examples": 289521, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 118856, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-cy-en-v1.0.tar.gz": {"num_bytes": 6487005, "checksum": "935145550ed2fbef59ba947b9a15d5281039fbbc5377624af3011ed4d454f6b0"}}, "download_size": 6487005, "post_processing_size": null, "dataset_size": 17488125, "size_in_bytes": 23975130}, "da-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["da", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "da", "output": "en"}, "builder_name": "opus100", "config_name": "da-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 298123, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 126425274, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 300624, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-da-en-v1.0.tar.gz": {"num_bytes": 50404122, "checksum": "20fcc66a488897a549495a11427a9f3b84926a323ddaa0372f03b14fabee4f14"}}, "download_size": 50404122, "post_processing_size": null, "dataset_size": 127024021, "size_in_bytes": 177428143}, "de-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["de", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "de", "output": "en"}, "builder_name": "opus100", "config_name": "de-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 330959, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 152246756, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 332350, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-de-en-v1.0.tar.gz": {"num_bytes": 67205361, "checksum": "aca2cb2f71268700c5734c7b177fbc91dcadc1ac59844000a85d9d10d0bb19c0"}}, "download_size": 67205361, "post_processing_size": null, "dataset_size": 152910065, "size_in_bytes": 220115426}, "dz-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["dz", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "dz", "output": "en"}, "builder_name": "opus100", "config_name": "dz-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 81162, "num_examples": 624, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-dz-en-v1.0.tar.gz": {"num_bytes": 17814, "checksum": "c873dfd45a93420ac96bdd05d2835a7ee57bceff7f2258cfda8aecb2787bbe05"}}, "download_size": 17814, "post_processing_size": null, "dataset_size": 81162, "size_in_bytes": 98976}, "el-en": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["el", "en"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "el", "output": "en"}, "builder_name": "opus100", "config_name": "el-en", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 302393, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 127964703, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 291234, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-el-en-v1.0.tar.gz": {"num_bytes": 43973686, "checksum": "2992dd7a8cadd2c3db7b7267dd3df4cc5f3dee21a2514d6218a718891f543f99"}}, "download_size": 43973686, "post_processing_size": null, "dataset_size": 128558330, "size_in_bytes": 172532016}, "en-eo": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "eo"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "eo"}, "builder_name": "opus100", "config_name": "en-eo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 167386, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 24431953, "num_examples": 337106, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 168838, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-eo-v1.0.tar.gz": {"num_bytes": 9999313, "checksum": "1db2518821cac8d313085f2a79dc3dcaa30ca35dab8436786e70921209c52a87"}}, "download_size": 9999313, "post_processing_size": null, "dataset_size": 24768177, "size_in_bytes": 34767490}, "en-es": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "es"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "es"}, "builder_name": "opus100", "config_name": "en-es", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 326270, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 136643904, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 326735, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-es-v1.0.tar.gz": {"num_bytes": 55534068, "checksum": "e361cd112e9831d50e48d59fa0a9630a51daf323ed6e555fa25464b5e409c649"}}, "download_size": 55534068, "post_processing_size": null, "dataset_size": 137296909, "size_in_bytes": 192830977}, "en-et": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "et"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "et"}, "builder_name": "opus100", "config_name": "en-et", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 272171, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 112299053, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 276962, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-et-v1.0.tar.gz": {"num_bytes": 46235623, "checksum": "9fefd0239214e3620b8e525c7bab5ddbac4b8cdcdaf784936be9cedde5cef421"}}, "download_size": 46235623, "post_processing_size": null, "dataset_size": 112848186, "size_in_bytes": 159083809}, "en-eu": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "eu"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "eu"}, "builder_name": "opus100", "config_name": "en-eu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 280885, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 112330085, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 281503, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-eu-v1.0.tar.gz": {"num_bytes": 46389313, "checksum": "b280ab292fad2b0c20599b0c0032eb1232706327228cdfe91611bbcf64979b9a"}}, "download_size": 46389313, "post_processing_size": null, "dataset_size": 112892473, "size_in_bytes": 159281786}, "en-fa": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "fa"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "fa"}, "builder_name": "opus100", "config_name": "en-fa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 296556, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 125401335, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 291129, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-fa-v1.0.tar.gz": {"num_bytes": 44568447, "checksum": "000305cca8a8c6278f8d036c06845f4c994d1343f4710873f99b705aeb529e51"}}, "download_size": 44568447, "post_processing_size": null, "dataset_size": 125989020, "size_in_bytes": 170557467}, "en-fi": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "fi"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "fi"}, "builder_name": "opus100", "config_name": "en-fi", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 245822, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 106025790, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 247227, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-fi-v1.0.tar.gz": {"num_bytes": 42563103, "checksum": "fa888ce283e2887a6932ab007518f2b2586e6d80ebe44856f8648e47e0c83b0f"}}, "download_size": 42563103, "post_processing_size": null, "dataset_size": 106518839, "size_in_bytes": 149081942}, "en-fr": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "fr"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "fr"}, "builder_name": "opus100", "config_name": "en-fr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 469731, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 201441250, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 481484, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-fr-v1.0.tar.gz": {"num_bytes": 81009778, "checksum": "c5521d61d7a158bcdebdd590c86b6e26e7d434cb1e02ca12681f8f82ed998c2d"}}, "download_size": 81009778, "post_processing_size": null, "dataset_size": 202392465, "size_in_bytes": 283402243}, "en-fy": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "fy"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "fy"}, "builder_name": "opus100", "config_name": "en-fy", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 101246, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 3895688, "num_examples": 54342, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 100129, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-fy-v1.0.tar.gz": {"num_bytes": 1522187, "checksum": "bc86f0ee0e99d4fd9e2bb3c1c27ce8f50ee0ca56fe12b75e3d8d4abb97a267eb"}}, "download_size": 1522187, "post_processing_size": null, "dataset_size": 4097063, "size_in_bytes": 5619250}, "en-ga": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ga"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ga"}, "builder_name": "opus100", "config_name": "en-ga", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 503317, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 42132742, "num_examples": 289524, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 503217, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ga-v1.0.tar.gz": {"num_bytes": 14998873, "checksum": "8f278d5c9cdc9e4ce660a9afe72a286ef4651adfb9f4d2fe0f38b4cacb8be449"}}, "download_size": 14998873, "post_processing_size": null, "dataset_size": 43139276, "size_in_bytes": 58138149}, "en-gd": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "gd"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "gd"}, "builder_name": "opus100", "config_name": "en-gd", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 218362, "num_examples": 1606, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 1254795, "num_examples": 16316, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 203885, "num_examples": 1605, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-gd-v1.0.tar.gz": {"num_bytes": 564053, "checksum": "fc567fb3065c700046120bbdcd3b5a54eb8f2cfe3d7a97dcad5270ff2e342343"}}, "download_size": 564053, "post_processing_size": null, "dataset_size": 1677042, "size_in_bytes": 2241095}, "en-gl": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "gl"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "gl"}, "builder_name": "opus100", "config_name": "en-gl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 190699, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 43327444, "num_examples": 515344, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 193606, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-gl-v1.0.tar.gz": {"num_bytes": 18056665, "checksum": "ff2e4ca48fd491a8bff5400a5584087ba05c893778845f2167c218ad2d57f4a3"}}, "download_size": 18056665, "post_processing_size": null, "dataset_size": 43711749, "size_in_bytes": 61768414}, "en-gu": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "gu"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "gu"}, "builder_name": "opus100", "config_name": "en-gu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 199733, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 33641975, "num_examples": 318306, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 205550, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-gu-v1.0.tar.gz": {"num_bytes": 9407543, "checksum": "ea614ac05ca527f903604603cf674b2825058e12a1bfb74eba40c77d52e773f5"}}, "download_size": 9407543, "post_processing_size": null, "dataset_size": 34047258, "size_in_bytes": 43454801}, "en-ha": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ha"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ha"}, "builder_name": "opus100", "config_name": "en-ha", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 407352, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 20391964, "num_examples": 97983, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 411526, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ha-v1.0.tar.gz": {"num_bytes": 6898482, "checksum": "e2c24892c72a530e43eaea0e3033291b8ea166a2c8665955a074f56cacd69136"}}, "download_size": 6898482, "post_processing_size": null, "dataset_size": 21210842, "size_in_bytes": 28109324}, "en-he": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "he"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "he"}, "builder_name": "opus100", "config_name": "en-he", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 208475, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 91160431, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 209446, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-he-v1.0.tar.gz": {"num_bytes": 31214136, "checksum": "693831c44c3afc5f34d7e3881dddd9eea6ba9ad5ec08d63359f4ab9bb033a758"}}, "download_size": 31214136, "post_processing_size": null, "dataset_size": 91578352, "size_in_bytes": 122792488}, "en-hi": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "hi"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "hi"}, "builder_name": "opus100", "config_name": "en-hi", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 496578, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 124923977, "num_examples": 534319, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 474087, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-hi-v1.0.tar.gz": {"num_bytes": 35993452, "checksum": "f6efc00fd95972b9b13157167356152766d962db07b6e73e6cc87479083c3b90"}}, "download_size": 35993452, "post_processing_size": null, "dataset_size": 125894642, "size_in_bytes": 161888094}, "en-hr": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "hr"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "hr"}, "builder_name": "opus100", "config_name": "en-hr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 179644, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 75310316, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 179623, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-hr-v1.0.tar.gz": {"num_bytes": 30728154, "checksum": "7b2302e36e747700ed0036c119004d31a7955918af7e1f11dc868e81c4c9f537"}}, "download_size": 30728154, "post_processing_size": null, "dataset_size": 75669583, "size_in_bytes": 106397737}, "en-hu": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "hu"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "hu"}, "builder_name": "opus100", "config_name": "en-hu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 206047, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 87484262, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 208315, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-hu-v1.0.tar.gz": {"num_bytes": 35696235, "checksum": "04b51fd88964c0a42f49dc779e1d515944872ae1a998a6d3602cf63aa4d2d216"}}, "download_size": 35696235, "post_processing_size": null, "dataset_size": 87898624, "size_in_bytes": 123594859}, "en-hy": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "hy"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "hy"}, "builder_name": "opus100", "config_name": "en-hy", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 652631, "num_examples": 7059, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-hy-v1.0.tar.gz": {"num_bytes": 215246, "checksum": "dbcd452a3a0eb3c1485621ea314a096ab5ddf89e84e1f02fe7b4bdf4cc268dcc"}}, "download_size": 215246, "post_processing_size": null, "dataset_size": 652631, "size_in_bytes": 867877}, "en-id": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "id"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "id"}, "builder_name": "opus100", "config_name": "en-id", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 177693, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 78699773, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 180032, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-id-v1.0.tar.gz": {"num_bytes": 29914089, "checksum": "3466dd5f90b0e44542edafac932c25fbeabcad65b5cd50d264319f1fd18c1b8d"}}, "download_size": 29914089, "post_processing_size": null, "dataset_size": 79057498, "size_in_bytes": 108971587}, "en-ig": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ig"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ig"}, "builder_name": "opus100", "config_name": "en-ig", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 137332, "num_examples": 1843, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 1612539, "num_examples": 18415, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 135995, "num_examples": 1843, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ig-v1.0.tar.gz": {"num_bytes": 391849, "checksum": "4cbee7925658026f1791d2638edbc1f72435306f39428f76627fa84506142755"}}, "download_size": 391849, "post_processing_size": null, "dataset_size": 1885866, "size_in_bytes": 2277715}, "en-is": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "is"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "is"}, "builder_name": "opus100", "config_name": "en-is", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 170887, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 73964915, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 170640, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-is-v1.0.tar.gz": {"num_bytes": 28831218, "checksum": "1dfc59b45631dbf849c75a23b14379b05296cf651e24dd2df7ab3ab85f916624"}}, "download_size": 28831218, "post_processing_size": null, "dataset_size": 74306442, "size_in_bytes": 103137660}, "en-it": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "it"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "it"}, "builder_name": "opus100", "config_name": "en-it", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 299037, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 123655086, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 294362, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-it-v1.0.tar.gz": {"num_bytes": 50903618, "checksum": "54da7db3f7bcbd21e8b41180551d4fd5de0eb1802bfb2ec2b5d44bb919ba464a"}}, "download_size": 50903618, "post_processing_size": null, "dataset_size": 124248485, "size_in_bytes": 175152103}, "en-ja": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ja"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ja"}, "builder_name": "opus100", "config_name": "en-ja", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 190999, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 88349369, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 191419, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ja-v1.0.tar.gz": {"num_bytes": 34452575, "checksum": "78e83ee1b0e17834a02c2f51df1ce71d455b9211e46f113b97d441e40fed0681"}}, "download_size": 34452575, "post_processing_size": null, "dataset_size": 88731787, "size_in_bytes": 123184362}, "en-ka": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ka"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ka"}, "builder_name": "opus100", "config_name": "en-ka", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 256227, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 42465706, "num_examples": 377306, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 260416, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ka-v1.0.tar.gz": {"num_bytes": 12743188, "checksum": "6b189050fad96016e21a24869b8164da7450214197c473584ed1445bf626f617"}}, "download_size": 12743188, "post_processing_size": null, "dataset_size": 42982349, "size_in_bytes": 55725537}, "en-kk": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "kk"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "kk"}, "builder_name": "opus100", "config_name": "en-kk", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 137664, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 7124378, "num_examples": 79927, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 139665, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-kk-v1.0.tar.gz": {"num_bytes": 2425372, "checksum": "be76b8a0d8a60809cbb90320e7d484501c86686eb65c5aac585ad94b979da850"}}, "download_size": 2425372, "post_processing_size": null, "dataset_size": 7401707, "size_in_bytes": 9827079}, "en-km": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "km"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "km"}, "builder_name": "opus100", "config_name": "en-km", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 289027, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 19680611, "num_examples": 111483, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 302527, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-km-v1.0.tar.gz": {"num_bytes": 5193620, "checksum": "b65cbbf241a1eeeafd6b39c366f219356e6ce6e065d146a1e847200f055dd39b"}}, "download_size": 5193620, "post_processing_size": null, "dataset_size": 20272165, "size_in_bytes": 25465785}, "en-ko": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ko"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ko"}, "builder_name": "opus100", "config_name": "en-ko", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 190696, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 93665332, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 189368, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ko-v1.0.tar.gz": {"num_bytes": 37602794, "checksum": "f9040eb06807cf1cb2c0d3352249b38546bdbff668c242b98c95bfae84190fd2"}}, "download_size": 37602794, "post_processing_size": null, "dataset_size": 94045396, "size_in_bytes": 131648190}, "en-kn": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "kn"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "kn"}, "builder_name": "opus100", "config_name": "en-kn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 77205, "num_examples": 918, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 1833334, "num_examples": 14537, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 77607, "num_examples": 917, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-kn-v1.0.tar.gz": {"num_bytes": 525449, "checksum": "5c100de50142e36e750e6a0d595aeedccda7f1479d781b4a4a337c8361bfe87a"}}, "download_size": 525449, "post_processing_size": null, "dataset_size": 1988146, "size_in_bytes": 2513595}, "en-ku": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ku"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ku"}, "builder_name": "opus100", "config_name": "en-ku", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 247847, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 49107864, "num_examples": 144844, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 239325, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ku-v1.0.tar.gz": {"num_bytes": 14252198, "checksum": "b198288af58893bcb19e999565eb0180d100505003d5064643f51329024bf117"}}, "download_size": 14252198, "post_processing_size": null, "dataset_size": 49595036, "size_in_bytes": 63847234}, "en-ky": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ky"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ky"}, "builder_name": "opus100", "config_name": "en-ky", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 142530, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 1879298, "num_examples": 27215, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 138487, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ky-v1.0.tar.gz": {"num_bytes": 616902, "checksum": "2694ed66518822f283bfa8097e94160256e1dc9c405545258eb7962282ec62b4"}}, "download_size": 616902, "post_processing_size": null, "dataset_size": 2160315, "size_in_bytes": 2777217}, "en-li": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "li"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "li"}, "builder_name": "opus100", "config_name": "en-li", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 93350, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 1628601, "num_examples": 25535, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 92906, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-li-v1.0.tar.gz": {"num_bytes": 450092, "checksum": "e17e96b9331e90542e3a5145ae27e36a29a6433cbe1e4b6352bd6634546dfe7c"}}, "download_size": 450092, "post_processing_size": null, "dataset_size": 1814857, "size_in_bytes": 2264949}, "en-lt": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "lt"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "lt"}, "builder_name": "opus100", "config_name": "en-lt", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 482615, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 177061044, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 469117, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-lt-v1.0.tar.gz": {"num_bytes": 69388131, "checksum": "af798e0d92f37805bc9a8f3a13ff48a1e016b67938062016558f3a7cc39f8ded"}}, "download_size": 69388131, "post_processing_size": null, "dataset_size": 178012776, "size_in_bytes": 247400907}, "en-lv": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "lv"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "lv"}, "builder_name": "opus100", "config_name": "en-lv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 536576, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 206051849, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 522072, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-lv-v1.0.tar.gz": {"num_bytes": 78952903, "checksum": "5b824a15347a139655a8e630a724bc6df2d7d454d35b20da028630aa4969bb53"}}, "download_size": 78952903, "post_processing_size": null, "dataset_size": 207110497, "size_in_bytes": 286063400}, "en-mg": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "mg"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "mg"}, "builder_name": "opus100", "config_name": "en-mg", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 525067, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 130865649, "num_examples": 590771, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 511171, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-mg-v1.0.tar.gz": {"num_bytes": 52470504, "checksum": "b6f7242fd0d535422ca8958fdf295808ef5a47bb3f9901d26dc08ff2a8cbb843"}}, "download_size": 52470504, "post_processing_size": null, "dataset_size": 131901887, "size_in_bytes": 184372391}, "en-mk": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "mk"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "mk"}, "builder_name": "opus100", "config_name": "en-mk", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 308934, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 117069489, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 305498, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-mk-v1.0.tar.gz": {"num_bytes": 39517761, "checksum": "2fdb6bbfb189dbd6db5cea247488b26bf0d943cb82415a5f4cbe2d664d964fb1"}}, "download_size": 39517761, "post_processing_size": null, "dataset_size": 117683921, "size_in_bytes": 157201682}, "en-ml": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ml"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ml"}, "builder_name": "opus100", "config_name": "en-ml", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 340626, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 199971743, "num_examples": 822746, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 334459, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ml-v1.0.tar.gz": {"num_bytes": 48654808, "checksum": "0121284cdb5582b95aa344de6fd31322202da58e938763037e5cbf40d39e7644"}}, "download_size": 48654808, "post_processing_size": null, "dataset_size": 200646828, "size_in_bytes": 249301636}, "en-mn": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "mn"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "mn"}, "builder_name": "opus100", "config_name": "en-mn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 250778, "num_examples": 4294, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-mn-v1.0.tar.gz": {"num_bytes": 42039, "checksum": "52cbcf55978c5cc6be6062a3669edeb0c4833e0e8cdc235295a6c011b4fdaf7b"}}, "download_size": 42039, "post_processing_size": null, "dataset_size": 250778, "size_in_bytes": 292817}, "en-mr": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "mr"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "mr"}, "builder_name": "opus100", "config_name": "en-mr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 238612, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 2724131, "num_examples": 27007, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 235540, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-mr-v1.0.tar.gz": {"num_bytes": 910211, "checksum": "0988cfa9bac2da79b15c2a7a1cde0e31851b0d418c1edf3a0027c9b6fa6d43c2"}}, "download_size": 910211, "post_processing_size": null, "dataset_size": 3198283, "size_in_bytes": 4108494}, "en-ms": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ms"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ms"}, "builder_name": "opus100", "config_name": "en-ms", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 179705, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 76829645, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 180183, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ms-v1.0.tar.gz": {"num_bytes": 29807607, "checksum": "9f46a5e7dee938b2f9e51182b0d7939a3cea61c5cdda6d4fc0ba724d99688cd4"}}, "download_size": 29807607, "post_processing_size": null, "dataset_size": 77189533, "size_in_bytes": 106997140}, "en-mt": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "mt"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "mt"}, "builder_name": "opus100", "config_name": "en-mt", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 566134, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 222222396, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 594386, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-mt-v1.0.tar.gz": {"num_bytes": 84757608, "checksum": "327f2fbab86f4aec9df76c7e39d3f32b8f56b97877df1579f73c10f431f7305e"}}, "download_size": 84757608, "post_processing_size": null, "dataset_size": 223382916, "size_in_bytes": 308140524}, "en-my": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "my"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "my"}, "builder_name": "opus100", "config_name": "en-my", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 337351, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 3673501, "num_examples": 24594, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 336155, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-my-v1.0.tar.gz": {"num_bytes": 1038600, "checksum": "5338dfdbdea5a6178f4b3654e4f8644906d7c6d9b102d7e9c3631b65ab9debc9"}}, "download_size": 1038600, "post_processing_size": null, "dataset_size": 4347007, "size_in_bytes": 5385607}, "en-nb": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "nb"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "nb"}, "builder_name": "opus100", "config_name": "en-nb", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 334117, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 13611709, "num_examples": 142906, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 324400, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-nb-v1.0.tar.gz": {"num_bytes": 5706626, "checksum": "f184509d7a781fe60aaf617ca23a373488c369ca8eb6d5c2eebf7c74e747d6a7"}}, "download_size": 5706626, "post_processing_size": null, "dataset_size": 14270226, "size_in_bytes": 19976852}, "en-ne": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ne"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ne"}, "builder_name": "opus100", "config_name": "en-ne", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 186527, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 44136280, "num_examples": 406381, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 204920, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ne-v1.0.tar.gz": {"num_bytes": 11711988, "checksum": "0939f62d42db0f070c309b0fb02c4733e3e943cdbc504739fc0e4b42cbcd3ba8"}}, "download_size": 11711988, "post_processing_size": null, "dataset_size": 44527727, "size_in_bytes": 56239715}, "en-nl": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "nl"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "nl"}, "builder_name": "opus100", "config_name": "en-nl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 282755, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 112327073, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 270940, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-nl-v1.0.tar.gz": {"num_bytes": 45374708, "checksum": "8d59b586929585de9aa0f9805bc9873c5b9ea5075d34bf23ffb2714b02499308"}}, "download_size": 45374708, "post_processing_size": null, "dataset_size": 112880768, "size_in_bytes": 158255476}, "en-nn": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "nn"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "nn"}, "builder_name": "opus100", "config_name": "en-nn", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 179007, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 32924821, "num_examples": 486055, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 187650, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-nn-v1.0.tar.gz": {"num_bytes": 12742134, "checksum": "31c1fde5fb4b95c37cd9ceafcaf6104574727dc31b15d4b255f0ab71df4c6579"}}, "download_size": 12742134, "post_processing_size": null, "dataset_size": 33291478, "size_in_bytes": 46033612}, "en-no": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "no"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "no"}, "builder_name": "opus100", "config_name": "en-no", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 173328, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 74106283, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 178013, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-no-v1.0.tar.gz": {"num_bytes": 28851262, "checksum": "a8b20ca8dfd25581b54303a780c19105eab3645ad7981a1a821e93d634bfc4f6"}}, "download_size": 28851262, "post_processing_size": null, "dataset_size": 74457624, "size_in_bytes": 103308886}, "en-oc": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "oc"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "oc"}, "builder_name": "opus100", "config_name": "en-oc", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 82350, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 1627206, "num_examples": 35791, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 81650, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-oc-v1.0.tar.gz": {"num_bytes": 607192, "checksum": "9fe531860f1a7f89d97053fc4144b71fdbf6e734dc2d3d52dbc20714493cd0d4"}}, "download_size": 607192, "post_processing_size": null, "dataset_size": 1791206, "size_in_bytes": 2398398}, "en-or": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "or"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "or"}, "builder_name": "opus100", "config_name": "en-or", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 163947, "num_examples": 1318, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 1500749, "num_examples": 14273, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 155331, "num_examples": 1317, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-or-v1.0.tar.gz": {"num_bytes": 499401, "checksum": "dc15a00dbb19541ae46491490184d08302095faf7ed88d962f45cec768d87b05"}}, "download_size": 499401, "post_processing_size": null, "dataset_size": 1820027, "size_in_bytes": 2319428}, "en-pa": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "pa"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "pa"}, "builder_name": "opus100", "config_name": "en-pa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 133909, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 8509228, "num_examples": 107296, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 136196, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-pa-v1.0.tar.gz": {"num_bytes": 2589682, "checksum": "1107d5ae7c5b5c6a8575bce896cecb1e2b45b7dc454294222062245b6fe3f947"}}, "download_size": 2589682, "post_processing_size": null, "dataset_size": 8779333, "size_in_bytes": 11369015}, "en-pl": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "pl"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "pl"}, "builder_name": "opus100", "config_name": "en-pl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 212503, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 95248523, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 218216, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-pl-v1.0.tar.gz": {"num_bytes": 39320454, "checksum": "8684ef187956bfab33896f84437722e2c41ea00341eee7096039f0e39c4f2ddc"}}, "download_size": 39320454, "post_processing_size": null, "dataset_size": 95679242, "size_in_bytes": 134999696}, "en-ps": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ps"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ps"}, "builder_name": "opus100", "config_name": "en-ps", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 93003, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 4436576, "num_examples": 79127, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 95164, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ps-v1.0.tar.gz": {"num_bytes": 1223087, "checksum": "591b83f63a43b534653523681ac880bcaea2ff69c8f266112c66223bf44e90f5"}}, "download_size": 1223087, "post_processing_size": null, "dataset_size": 4624743, "size_in_bytes": 5847830}, "en-pt": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "pt"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "pt"}, "builder_name": "opus100", "config_name": "en-pt", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 296122, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 118243649, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 292082, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-pt-v1.0.tar.gz": {"num_bytes": 48087550, "checksum": "550e3382309dee492c77f8ce88c32fb559211ad4f032dca2d34c10ec0b0fe75d"}}, "download_size": 48087550, "post_processing_size": null, "dataset_size": 118831853, "size_in_bytes": 166919403}, "en-ro": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ro"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ro"}, "builder_name": "opus100", "config_name": "en-ro", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 198647, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 85249851, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 199172, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ro-v1.0.tar.gz": {"num_bytes": 35032743, "checksum": "4c87e36f2ff5eae5ffefb66abcf6c65a5f8af673946c911f10ab42bc468c5d42"}}, "download_size": 35032743, "post_processing_size": null, "dataset_size": 85647670, "size_in_bytes": 120680413}, "en-ru": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ru"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ru"}, "builder_name": "opus100", "config_name": "en-ru", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 490984, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 195101737, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 490246, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ru-v1.0.tar.gz": {"num_bytes": 68501634, "checksum": "422c96c952e37a0e7ae924538f6c521a9dcedb200e0b29eab07a6e3fd00cca9d"}}, "download_size": 68501634, "post_processing_size": null, "dataset_size": 196082967, "size_in_bytes": 264584601}, "en-rw": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "rw"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "rw"}, "builder_name": "opus100", "config_name": "en-rw", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 136197, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 15286303, "num_examples": 173823, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 134965, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-rw-v1.0.tar.gz": {"num_bytes": 5233241, "checksum": "ac65b6ee8d4900c078a53d90af463952b1f2f67333f36ccd53bc34cafa424ee4"}}, "download_size": 5233241, "post_processing_size": null, "dataset_size": 15557465, "size_in_bytes": 20790706}, "en-se": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "se"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "se"}, "builder_name": "opus100", "config_name": "en-se", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 85705, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 2047412, "num_examples": 35907, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 83672, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-se-v1.0.tar.gz": {"num_bytes": 806982, "checksum": "ccc109d695c8d238f5d8c318b1a329b4f4bd434452bd178c0db90e7df60bae0a"}}, "download_size": 806982, "post_processing_size": null, "dataset_size": 2216789, "size_in_bytes": 3023771}, "en-sh": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "sh"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "sh"}, "builder_name": "opus100", "config_name": "en-sh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 569487, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 60900239, "num_examples": 267211, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 555602, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-sh-v1.0.tar.gz": {"num_bytes": 22357505, "checksum": "68018403e648b55cca750e13d7c91cf6ec12be3a3cabcf83247200150fa39594"}}, "download_size": 22357505, "post_processing_size": null, "dataset_size": 62025328, "size_in_bytes": 84382833}, "en-si": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "si"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "si"}, "builder_name": "opus100", "config_name": "en-si", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 271743, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 114951675, "num_examples": 979109, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 271244, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-si-v1.0.tar.gz": {"num_bytes": 33247484, "checksum": "2a654e11df7040ab70f36cde1d84a431696cb834c5d25abe0bfbd375615262a6"}}, "download_size": 33247484, "post_processing_size": null, "dataset_size": 115494662, "size_in_bytes": 148742146}, "en-sk": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "sk"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "sk"}, "builder_name": "opus100", "config_name": "en-sk", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 258042, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 111743868, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 255470, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-sk-v1.0.tar.gz": {"num_bytes": 46618395, "checksum": "1641ebabc596e1dad84ec1bfccfdc0d8c65837fc8739631eb7826066bb328f05"}}, "download_size": 46618395, "post_processing_size": null, "dataset_size": 112257380, "size_in_bytes": 158875775}, "en-sl": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "sl"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "sl"}, "builder_name": "opus100", "config_name": "en-sl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 205478, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 90270957, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 198662, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-sl-v1.0.tar.gz": {"num_bytes": 37536724, "checksum": "57e5149d179d95eb7364220bcf10afbf419d94b402ad6dd8108610ae3e9e6760"}}, "download_size": 37536724, "post_processing_size": null, "dataset_size": 90675097, "size_in_bytes": 128211821}, "en-sq": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "sq"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "sq"}, "builder_name": "opus100", "config_name": "en-sq", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 275379, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 105745981, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 267312, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-sq-v1.0.tar.gz": {"num_bytes": 42697338, "checksum": "e2a9b220ad63824188297d184596a449c4a8c1ef3e1a681b63d46a9819c9ba62"}}, "download_size": 42697338, "post_processing_size": null, "dataset_size": 106288672, "size_in_bytes": 148986010}, "en-sr": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "sr"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "sr"}, "builder_name": "opus100", "config_name": "en-sr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 180232, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 75726835, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 184246, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-sr-v1.0.tar.gz": {"num_bytes": 31260575, "checksum": "a5660570f8dbb47f1c275bce55542f0e1e1a938645d24c7bdd01c941d5481b4f"}}, "download_size": 31260575, "post_processing_size": null, "dataset_size": 76091313, "size_in_bytes": 107351888}, "en-sv": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "sv"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "sv"}, "builder_name": "opus100", "config_name": "en-sv", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 271014, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 116985953, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 279994, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-sv-v1.0.tar.gz": {"num_bytes": 46694960, "checksum": "ab83d088b08bc1a9f7da7895b4caf4dd467a57b1be9096bc0b5c88d045156119"}}, "download_size": 46694960, "post_processing_size": null, "dataset_size": 117536961, "size_in_bytes": 164231921}, "en-ta": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ta"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ta"}, "builder_name": "opus100", "config_name": "en-ta", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 351990, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 74044524, "num_examples": 227014, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 335557, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ta-v1.0.tar.gz": {"num_bytes": 17652443, "checksum": "612322d09aa48b7875ee5eee94791aa5bc22ebd2b0411997d66682a0329b61d4"}}, "download_size": 17652443, "post_processing_size": null, "dataset_size": 74732071, "size_in_bytes": 92384514}, "en-te": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "te"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "te"}, "builder_name": "opus100", "config_name": "en-te", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 190595, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 6688625, "num_examples": 64352, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 193666, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-te-v1.0.tar.gz": {"num_bytes": 2011832, "checksum": "8bb0c2405cc62b37c1ec21584765e703b5739c2730c04898e6a18ff416c08529"}}, "download_size": 2011832, "post_processing_size": null, "dataset_size": 7072886, "size_in_bytes": 9084718}, "en-tg": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "tg"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "tg"}, "builder_name": "opus100", "config_name": "en-tg", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 372120, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 35477177, "num_examples": 193882, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 371728, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-tg-v1.0.tar.gz": {"num_bytes": 11389877, "checksum": "30caf792aa42db984dcc7704a09250b15a1a1b46a7d82e79d2c116ee4c7eeefa"}}, "download_size": 11389877, "post_processing_size": null, "dataset_size": 36221025, "size_in_bytes": 47610902}, "en-th": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "th"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "th"}, "builder_name": "opus100", "config_name": "en-th", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 290581, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 132821031, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 288366, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-th-v1.0.tar.gz": {"num_bytes": 38147204, "checksum": "223dc991bd45ff25e96dada988f97bedb0c6f28e930bd8ab3ff1344e4a2a5bb6"}}, "download_size": 38147204, "post_processing_size": null, "dataset_size": 133399978, "size_in_bytes": 171547182}, "en-tk": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "tk"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "tk"}, "builder_name": "opus100", "config_name": "en-tk", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 83886, "num_examples": 1852, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 719633, "num_examples": 13110, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 81014, "num_examples": 1852, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-tk-v1.0.tar.gz": {"num_bytes": 157481, "checksum": "afbbf17a9600d41176c522a760a54d1e143e7855063893934159cbda8e4e427b"}}, "download_size": 157481, "post_processing_size": null, "dataset_size": 884533, "size_in_bytes": 1042014}, "en-tr": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "tr"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "tr"}, "builder_name": "opus100", "config_name": "en-tr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 183833, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 78946365, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 181917, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-tr-v1.0.tar.gz": {"num_bytes": 30892429, "checksum": "0d4a941721721c94e99013052491894062d7782efb94a6593a91fd1f0f06cba0"}}, "download_size": 30892429, "post_processing_size": null, "dataset_size": 79312115, "size_in_bytes": 110204544}, "en-tt": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "tt"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "tt"}, "builder_name": "opus100", "config_name": "en-tt", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 693276, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 35313258, "num_examples": 100843, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 701670, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-tt-v1.0.tar.gz": {"num_bytes": 9940523, "checksum": "d23d7043364d55f5cf47abf93193a831bfdc5c0ec38052ac84bdb166b98348fb"}}, "download_size": 9940523, "post_processing_size": null, "dataset_size": 36708204, "size_in_bytes": 46648727}, "en-ug": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ug"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ug"}, "builder_name": "opus100", "config_name": "en-ug", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 620881, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 31576580, "num_examples": 72170, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 631236, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ug-v1.0.tar.gz": {"num_bytes": 8687743, "checksum": "176ccf6f66439cda78892db6c0f8bad67fc3f9ebe25feac7cd405ecff2d785d7"}}, "download_size": 8687743, "post_processing_size": null, "dataset_size": 32828697, "size_in_bytes": 41516440}, "en-uk": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "uk"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "uk"}, "builder_name": "opus100", "config_name": "en-uk", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 249750, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 104230356, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 247131, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-uk-v1.0.tar.gz": {"num_bytes": 37415496, "checksum": "09a689c33567da03908d134d240c6e353319408ada4fddd949e46bc7d2a675ea"}}, "download_size": 37415496, "post_processing_size": null, "dataset_size": 104727237, "size_in_bytes": 142142733}, "en-ur": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "ur"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "ur"}, "builder_name": "opus100", "config_name": "en-ur", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 538564, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 268961304, "num_examples": 753913, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 529316, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-ur-v1.0.tar.gz": {"num_bytes": 81092186, "checksum": "5e6dfe7bb1c956b37594a2b81213dde5f7dba1c0e49ecccdebcb4aa73af23711"}}, "download_size": 81092186, "post_processing_size": null, "dataset_size": 270029184, "size_in_bytes": 351121370}, "en-uz": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "uz"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "uz"}, "builder_name": "opus100", "config_name": "en-uz", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 408683, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 38375434, "num_examples": 173157, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 398861, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-uz-v1.0.tar.gz": {"num_bytes": 11791643, "checksum": "364902a0e71c5e7b3dcb6bae4517fd094a445cad02566d81889222b06bed2650"}}, "download_size": 11791643, "post_processing_size": null, "dataset_size": 39182978, "size_in_bytes": 50974621}, "en-vi": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "vi"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "vi"}, "builder_name": "opus100", "config_name": "en-vi", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 192752, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 82615270, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 194729, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-vi-v1.0.tar.gz": {"num_bytes": 30647296, "checksum": "6a37b14536cfa554ea90f8560e3f93e5273c5783e062fc07bb2e3ffbc07daf95"}}, "download_size": 30647296, "post_processing_size": null, "dataset_size": 83002751, "size_in_bytes": 113650047}, "en-wa": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "wa"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "wa"}, "builder_name": "opus100", "config_name": "en-wa", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 87099, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 6085948, "num_examples": 104496, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 87726, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-wa-v1.0.tar.gz": {"num_bytes": 2119821, "checksum": "5b60c912eb04afe15df000ca3f3e01c33a63ced61062bb1bb79d007a78bb021d"}}, "download_size": 2119821, "post_processing_size": null, "dataset_size": 6260773, "size_in_bytes": 8380594}, "en-xh": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "xh"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "xh"}, "builder_name": "opus100", "config_name": "en-xh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 318660, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 50607248, "num_examples": 439671, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 315839, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-xh-v1.0.tar.gz": {"num_bytes": 20503199, "checksum": "89030efe5382cce5c1c4085651fad222e04f43487ac1082b7aeeed8b6f8e2bf4"}}, "download_size": 20503199, "post_processing_size": null, "dataset_size": 51241747, "size_in_bytes": 71744946}, "en-yi": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "yi"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "yi"}, "builder_name": "opus100", "config_name": "en-yi", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 96490, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 1275143, "num_examples": 15010, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 99826, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-yi-v1.0.tar.gz": {"num_bytes": 284031, "checksum": "d064b552fedb167dcdf9979db338e4c239cabbe238fd03755d61775c4dbb3945"}}, "download_size": 284031, "post_processing_size": null, "dataset_size": 1471459, "size_in_bytes": 1755490}, "en-yo": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "yo"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "yo"}, "builder_name": "opus100", "config_name": "en-yo", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 979769, "num_examples": 10375, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-yo-v1.0.tar.gz": {"num_bytes": 177540, "checksum": "42a74826c0c29b35255186a706c3b15b43995d9ca34b0e68aa60700be9069e06"}}, "download_size": 177540, "post_processing_size": null, "dataset_size": 979769, "size_in_bytes": 1157309}, "en-zh": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "zh"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "zh"}, "builder_name": "opus100", "config_name": "en-zh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 511372, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 200062983, "num_examples": 1000000, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 512364, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-zh-v1.0.tar.gz": {"num_bytes": 83265500, "checksum": "c417af397ec380032528a9de8e2fbed40e2de78e0bcc6bace0de257a01ae6635"}}, "download_size": 83265500, "post_processing_size": null, "dataset_size": 201086719, "size_in_bytes": 284352219}, "en-zu": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["en", "zu"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "en", "output": "zu"}, "builder_name": "opus100", "config_name": "en-zu", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 117518, "num_examples": 2000, "dataset_name": "opus100"}, "train": {"name": "train", "num_bytes": 2799590, "num_examples": 38616, "dataset_name": "opus100"}, "validation": {"name": "validation", "num_bytes": 120141, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-en-zu-v1.0.tar.gz": {"num_bytes": 889951, "checksum": "0b488dd9205d2801b80ae9e72a946ef1f8ecd4dbe47c54c8788c8c62b7397056"}}, "download_size": 889951, "post_processing_size": null, "dataset_size": 3037249, "size_in_bytes": 3927200}, "ar-de": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["ar", "de"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "ar", "output": "de"}, "builder_name": "opus100", "config_name": "ar-de", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 238599, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 238599, "size_in_bytes": 2795390}, "ar-fr": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["ar", "fr"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "ar", "output": "fr"}, "builder_name": "opus100", "config_name": "ar-fr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 547382, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 547382, "size_in_bytes": 3104173}, "ar-nl": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["ar", "nl"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "ar", "output": "nl"}, "builder_name": "opus100", "config_name": "ar-nl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 212936, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 212936, "size_in_bytes": 2769727}, "ar-ru": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["ar", "ru"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "ar", "output": "ru"}, "builder_name": "opus100", "config_name": "ar-ru", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 808270, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 808270, "size_in_bytes": 3365061}, "ar-zh": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["ar", "zh"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "ar", "output": "zh"}, "builder_name": "opus100", "config_name": "ar-zh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 713412, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 713412, "size_in_bytes": 3270203}, "de-fr": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["de", "fr"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "de", "output": "fr"}, "builder_name": "opus100", "config_name": "de-fr", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 458746, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 458746, "size_in_bytes": 3015537}, "de-nl": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["de", "nl"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "de", "output": "nl"}, "builder_name": "opus100", "config_name": "de-nl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 403886, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 403886, "size_in_bytes": 2960677}, "de-ru": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["de", "ru"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "de", "output": "ru"}, "builder_name": "opus100", "config_name": "de-ru", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 315779, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 315779, "size_in_bytes": 2872570}, "de-zh": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["de", "zh"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "de", "output": "zh"}, "builder_name": "opus100", "config_name": "de-zh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 280397, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 280397, "size_in_bytes": 2837188}, "fr-nl": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["fr", "nl"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "fr", "output": "nl"}, "builder_name": "opus100", "config_name": "fr-nl", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 368646, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 368646, "size_in_bytes": 2925437}, "fr-ru": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["fr", "ru"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "fr", "output": "ru"}, "builder_name": "opus100", "config_name": "fr-ru", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 732724, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 732724, "size_in_bytes": 3289515}, "fr-zh": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["fr", "zh"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "fr", "output": "zh"}, "builder_name": "opus100", "config_name": "fr-zh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 619394, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 619394, "size_in_bytes": 3176185}, "nl-ru": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["nl", "ru"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "nl", "output": "ru"}, "builder_name": "opus100", "config_name": "nl-ru", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 256067, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 256067, "size_in_bytes": 2812858}, "nl-zh": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["nl", "zh"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "nl", "output": "zh"}, "builder_name": "opus100", "config_name": "nl-zh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 183641, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 183641, "size_in_bytes": 2740432}, "ru-zh": {"description": "OPUS-100 is English-centric, meaning that all training pairs include English on either the source or target side.\nThe corpus covers 100 languages (including English).OPUS-100 contains approximately 55M sentence pairs.\nOf the 99 language pairs, 44 have 1M sentence pairs of training data, 73 have at least 100k, and 95 have at least 10k.\n", "citation": "@misc{zhang2020improving,\n title={Improving Massively Multilingual Neural Machine Translation and Zero-Shot Translation},\n author={Biao Zhang and Philip Williams and Ivan Titov and Rico Sennrich},\n year={2020},\n eprint={2004.11867},\n archivePrefix={arXiv},\n primaryClass={cs.CL}\n}\n", "homepage": "http://opus.nlpl.eu/opus-100.php", "license": "", "features": {"translation": {"languages": ["ru", "zh"], "id": null, "_type": "Translation"}}, "post_processed": null, "supervised_keys": {"input": "ru", "output": "zh"}, "builder_name": "opus100", "config_name": "ru-zh", "version": {"version_str": "0.0.0", "description": null, "major": 0, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 916114, "num_examples": 2000, "dataset_name": "opus100"}}, "download_checksums": {"https://object.pouta.csc.fi/OPUS-100/v1.0/opus-100-corpus-zeroshot-v1.0.tar.gz": {"num_bytes": 2556791, "checksum": "241de5bf4eb382017337f3ffb23cdf319c971585b5f679da9a420f23a5c2ed8f"}}, "download_size": 2556791, "post_processing_size": null, "dataset_size": 916114, "size_in_bytes": 3472905}}