File size: 155,512 Bytes
522773a |
1 |
{"de.100": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "de.100", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 137014, "num_examples": 1047, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 139422, "num_examples": 1013, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 203907, "num_examples": 1586, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 212932, "num_examples": 1582, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}}, "download_size": 3743216, "post_processing_size": null, "dataset_size": 693275, "size_in_bytes": 4436491}, "de.95": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "de.95", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 137014, "num_examples": 1047, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 139422, "num_examples": 1013, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 203907, "num_examples": 1586, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 212932, "num_examples": 1582, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 68749238, "num_examples": 590000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_de.70.jsonl.bz2": {"num_bytes": 81804580, "checksum": "4fadd1573aa7e372c0ab93e46d3ff22a249df3538a74d24b429693e575ffb995"}}, "download_size": 85547796, "post_processing_size": null, "dataset_size": 69442513, "size_in_bytes": 154990309}, "de.90": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "de.90", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 137014, "num_examples": 1047, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 139422, "num_examples": 1013, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 203907, "num_examples": 1586, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 212932, "num_examples": 1582, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 116327588, "num_examples": 1000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_de.70.jsonl.bz2": {"num_bytes": 81804580, "checksum": "4fadd1573aa7e372c0ab93e46d3ff22a249df3538a74d24b429693e575ffb995"}}, "download_size": 85547796, "post_processing_size": null, "dataset_size": 117020863, "size_in_bytes": 202568659}, "de.85": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "de.85", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 137014, "num_examples": 1047, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 139422, "num_examples": 1013, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 203907, "num_examples": 1586, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 212932, "num_examples": 1582, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 241666427, "num_examples": 2000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_de.70.jsonl.bz2": {"num_bytes": 81804580, "checksum": "4fadd1573aa7e372c0ab93e46d3ff22a249df3538a74d24b429693e575ffb995"}}, "download_size": 85547796, "post_processing_size": null, "dataset_size": 242359702, "size_in_bytes": 327907498}, "de.80": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "de.80", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 137014, "num_examples": 1047, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 139422, "num_examples": 1013, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 203907, "num_examples": 1586, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 212932, "num_examples": 1582, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 479785654, "num_examples": 3800000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_de.70.jsonl.bz2": {"num_bytes": 81804580, "checksum": "4fadd1573aa7e372c0ab93e46d3ff22a249df3538a74d24b429693e575ffb995"}}, "download_size": 85547796, "post_processing_size": null, "dataset_size": 480478929, "size_in_bytes": 566026725}, "de.75": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "de.75", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 137014, "num_examples": 1047, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 139422, "num_examples": 1013, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 203907, "num_examples": 1586, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 212932, "num_examples": 1582, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 598415516, "num_examples": 4700000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_de.70.jsonl.bz2": {"num_bytes": 81804580, "checksum": "4fadd1573aa7e372c0ab93e46d3ff22a249df3538a74d24b429693e575ffb995"}}, "download_size": 85547796, "post_processing_size": null, "dataset_size": 599108791, "size_in_bytes": 684656587}, "de.70": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "de.70", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 137014, "num_examples": 1047, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 139422, "num_examples": 1013, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 203907, "num_examples": 1586, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 212932, "num_examples": 1582, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 742627098, "num_examples": 5800000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_de.70.jsonl.bz2": {"num_bytes": 81804580, "checksum": "4fadd1573aa7e372c0ab93e46d3ff22a249df3538a74d24b429693e575ffb995"}}, "download_size": 85547796, "post_processing_size": null, "dataset_size": 743320373, "size_in_bytes": 828868169}, "de.65": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "de.65", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 137014, "num_examples": 1047, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 139422, "num_examples": 1013, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 203907, "num_examples": 1586, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 212932, "num_examples": 1582, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 986464596, "num_examples": 7700000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_de.60.jsonl.bz2": {"num_bytes": 169221021, "checksum": "362b713b286d4c73c55617447b096285dc40ea32c948f443a98d57fb8d5c2c87"}}, "download_size": 172964237, "post_processing_size": null, "dataset_size": 987157871, "size_in_bytes": 1160122108}, "de.60": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "de.60", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 137014, "num_examples": 1047, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 139422, "num_examples": 1013, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 203907, "num_examples": 1586, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 212932, "num_examples": 1582, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 1566809284, "num_examples": 12500000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_de.60.jsonl.bz2": {"num_bytes": 169221021, "checksum": "362b713b286d4c73c55617447b096285dc40ea32c948f443a98d57fb8d5c2c87"}}, "download_size": 172964237, "post_processing_size": null, "dataset_size": 1567502559, "size_in_bytes": 1740466796}, "en.100": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "en.100", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122729, "num_examples": 982, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 133432, "num_examples": 1015, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 180418, "num_examples": 1445, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 190131, "num_examples": 1455, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_de.60.jsonl.bz2": {"num_bytes": 169221021, "checksum": "362b713b286d4c73c55617447b096285dc40ea32c948f443a98d57fb8d5c2c87"}}, "download_size": 172964237, "post_processing_size": null, "dataset_size": 626710, "size_in_bytes": 173590947}, "en.95": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "en.95", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122729, "num_examples": 982, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 133432, "num_examples": 1015, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 180418, "num_examples": 1445, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 190131, "num_examples": 1455, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 113422570, "num_examples": 1000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_en.70.jsonl.bz2": {"num_bytes": 267287175, "checksum": "40b240ef8140f76a12b863f2be60d45d23d84e5df51f92d02a2021e44ef3aa63"}}, "download_size": 271030391, "post_processing_size": null, "dataset_size": 114049280, "size_in_bytes": 385079671}, "en.90": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "en.90", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122729, "num_examples": 982, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 133432, "num_examples": 1015, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 180418, "num_examples": 1445, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 190131, "num_examples": 1455, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 169856418, "num_examples": 1500000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_en.70.jsonl.bz2": {"num_bytes": 267287175, "checksum": "40b240ef8140f76a12b863f2be60d45d23d84e5df51f92d02a2021e44ef3aa63"}}, "download_size": 271030391, "post_processing_size": null, "dataset_size": 170483128, "size_in_bytes": 441513519}, "en.85": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "en.85", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122729, "num_examples": 982, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 133432, "num_examples": 1015, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 180418, "num_examples": 1445, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 190131, "num_examples": 1455, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 423496913, "num_examples": 3500000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_en.70.jsonl.bz2": {"num_bytes": 267287175, "checksum": "40b240ef8140f76a12b863f2be60d45d23d84e5df51f92d02a2021e44ef3aa63"}}, "download_size": 271030391, "post_processing_size": null, "dataset_size": 424123623, "size_in_bytes": 695154014}, "en.80": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "en.80", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122729, "num_examples": 982, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 133432, "num_examples": 1015, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 180418, "num_examples": 1445, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 190131, "num_examples": 1455, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 645125750, "num_examples": 5200000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_en.70.jsonl.bz2": {"num_bytes": 267287175, "checksum": "40b240ef8140f76a12b863f2be60d45d23d84e5df51f92d02a2021e44ef3aa63"}}, "download_size": 271030391, "post_processing_size": null, "dataset_size": 645752460, "size_in_bytes": 916782851}, "en.75": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "en.75", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122729, "num_examples": 982, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 133432, "num_examples": 1015, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 180418, "num_examples": 1445, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 190131, "num_examples": 1455, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 878730637, "num_examples": 7000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_en.70.jsonl.bz2": {"num_bytes": 267287175, "checksum": "40b240ef8140f76a12b863f2be60d45d23d84e5df51f92d02a2021e44ef3aa63"}}, "download_size": 271030391, "post_processing_size": null, "dataset_size": 879357347, "size_in_bytes": 1150387738}, "en.70": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "en.70", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122729, "num_examples": 982, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 133432, "num_examples": 1015, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 180418, "num_examples": 1445, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 190131, "num_examples": 1455, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 2508698468, "num_examples": 20000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_en.70.jsonl.bz2": {"num_bytes": 267287175, "checksum": "40b240ef8140f76a12b863f2be60d45d23d84e5df51f92d02a2021e44ef3aa63"}}, "download_size": 271030391, "post_processing_size": null, "dataset_size": 2509325178, "size_in_bytes": 2780355569}, "en.65": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "en.65", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122729, "num_examples": 982, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 133432, "num_examples": 1015, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 180418, "num_examples": 1445, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 190131, "num_examples": 1455, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 3659521628, "num_examples": 30000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_en.60.jsonl.bz2": {"num_bytes": 459303693, "checksum": "229250e3dbbda18544a462c1c5f8e26fc7712b3340f6786d1be3ca74c439c646"}}, "download_size": 463046909, "post_processing_size": null, "dataset_size": 3660148338, "size_in_bytes": 4123195247}, "en.60": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "en.60", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122729, "num_examples": 982, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 133432, "num_examples": 1015, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 180418, "num_examples": 1445, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 190131, "num_examples": 1455, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 4203882328, "num_examples": 35000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_en.60.jsonl.bz2": {"num_bytes": 459303693, "checksum": "229250e3dbbda18544a462c1c5f8e26fc7712b3340f6786d1be3ca74c439c646"}}, "download_size": 463046909, "post_processing_size": null, "dataset_size": 4204509038, "size_in_bytes": 4667555947}, "fi.100": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fi.100", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122828, "num_examples": 958, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 132420, "num_examples": 963, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 218462, "num_examples": 1749, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 232958, "num_examples": 1760, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_en.60.jsonl.bz2": {"num_bytes": 459303693, "checksum": "229250e3dbbda18544a462c1c5f8e26fc7712b3340f6786d1be3ca74c439c646"}}, "download_size": 463046909, "post_processing_size": null, "dataset_size": 706668, "size_in_bytes": 463753577}, "fi.95": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fi.95", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122828, "num_examples": 958, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 132420, "num_examples": 963, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 218462, "num_examples": 1749, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 232958, "num_examples": 1760, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 54700554, "num_examples": 480000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fi.70.jsonl.bz2": {"num_bytes": 51684779, "checksum": "67979a37a6aac2d833a1de1a36b751e6df8c98279b106b30f06560dfe52ebf8f"}}, "download_size": 55427995, "post_processing_size": null, "dataset_size": 55407222, "size_in_bytes": 110835217}, "fi.90": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fi.90", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122828, "num_examples": 958, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 132420, "num_examples": 963, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 218462, "num_examples": 1749, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 232958, "num_examples": 1760, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 73055381, "num_examples": 640000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fi.70.jsonl.bz2": {"num_bytes": 51684779, "checksum": "67979a37a6aac2d833a1de1a36b751e6df8c98279b106b30f06560dfe52ebf8f"}}, "download_size": 55427995, "post_processing_size": null, "dataset_size": 73762049, "size_in_bytes": 129190044}, "fi.85": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fi.85", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122828, "num_examples": 958, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 132420, "num_examples": 963, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 218462, "num_examples": 1749, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 232958, "num_examples": 1760, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 108258097, "num_examples": 950000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fi.70.jsonl.bz2": {"num_bytes": 51684779, "checksum": "67979a37a6aac2d833a1de1a36b751e6df8c98279b106b30f06560dfe52ebf8f"}}, "download_size": 55427995, "post_processing_size": null, "dataset_size": 108964765, "size_in_bytes": 164392760}, "fi.80": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fi.80", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122828, "num_examples": 958, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 132420, "num_examples": 963, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 218462, "num_examples": 1749, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 232958, "num_examples": 1760, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 146992910, "num_examples": 1300000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fi.70.jsonl.bz2": {"num_bytes": 51684779, "checksum": "67979a37a6aac2d833a1de1a36b751e6df8c98279b106b30f06560dfe52ebf8f"}}, "download_size": 55427995, "post_processing_size": null, "dataset_size": 147699578, "size_in_bytes": 203127573}, "fi.75": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fi.75", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122828, "num_examples": 958, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 132420, "num_examples": 963, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 218462, "num_examples": 1749, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 232958, "num_examples": 1760, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 360442479, "num_examples": 3000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fi.70.jsonl.bz2": {"num_bytes": 51684779, "checksum": "67979a37a6aac2d833a1de1a36b751e6df8c98279b106b30f06560dfe52ebf8f"}}, "download_size": 55427995, "post_processing_size": null, "dataset_size": 361149147, "size_in_bytes": 416577142}, "fi.70": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fi.70", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122828, "num_examples": 958, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 132420, "num_examples": 963, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 218462, "num_examples": 1749, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 232958, "num_examples": 1760, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 423492843, "num_examples": 3500000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fi.70.jsonl.bz2": {"num_bytes": 51684779, "checksum": "67979a37a6aac2d833a1de1a36b751e6df8c98279b106b30f06560dfe52ebf8f"}}, "download_size": 55427995, "post_processing_size": null, "dataset_size": 424199511, "size_in_bytes": 479627506}, "fi.65": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fi.65", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122828, "num_examples": 958, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 132420, "num_examples": 963, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 218462, "num_examples": 1749, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 232958, "num_examples": 1760, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 662127329, "num_examples": 5400000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fi.60.jsonl.bz2": {"num_bytes": 123097698, "checksum": "fedef68c6c7eadc4fd512035e637d094c33d4442f9ac6b0a5d94e180fe8d9701"}}, "download_size": 126840914, "post_processing_size": null, "dataset_size": 662833997, "size_in_bytes": 789674911}, "fi.60": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fi.60", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122828, "num_examples": 958, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 132420, "num_examples": 963, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 218462, "num_examples": 1749, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 232958, "num_examples": 1760, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 1084625934, "num_examples": 8900000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fi.60.jsonl.bz2": {"num_bytes": 123097698, "checksum": "fedef68c6c7eadc4fd512035e637d094c33d4442f9ac6b0a5d94e180fe8d9701"}}, "download_size": 126840914, "post_processing_size": null, "dataset_size": 1085332602, "size_in_bytes": 1212173516}, "fr.100": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fr.100", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 147905, "num_examples": 1007, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 155527, "num_examples": 997, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 231834, "num_examples": 1674, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 243452, "num_examples": 1630, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fi.60.jsonl.bz2": {"num_bytes": 123097698, "checksum": "fedef68c6c7eadc4fd512035e637d094c33d4442f9ac6b0a5d94e180fe8d9701"}}, "download_size": 126840914, "post_processing_size": null, "dataset_size": 778718, "size_in_bytes": 127619632}, "fr.95": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fr.95", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 147905, "num_examples": 1007, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 155527, "num_examples": 997, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 231834, "num_examples": 1674, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 243452, "num_examples": 1630, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 111339740, "num_examples": 940000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fr.70.jsonl.bz2": {"num_bytes": 183308200, "checksum": "9ae36c666dfb1e3a30fd9d33a2b5b2329f1ff498aef261b6a889e4767eb79b67"}}, "download_size": 187051416, "post_processing_size": null, "dataset_size": 112118458, "size_in_bytes": 299169874}, "fr.90": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fr.90", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 147905, "num_examples": 1007, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 155527, "num_examples": 997, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 231834, "num_examples": 1674, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 243452, "num_examples": 1630, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 292238564, "num_examples": 2400000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fr.70.jsonl.bz2": {"num_bytes": 183308200, "checksum": "9ae36c666dfb1e3a30fd9d33a2b5b2329f1ff498aef261b6a889e4767eb79b67"}}, "download_size": 187051416, "post_processing_size": null, "dataset_size": 293017282, "size_in_bytes": 480068698}, "fr.85": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fr.85", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 147905, "num_examples": 1007, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 155527, "num_examples": 997, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 231834, "num_examples": 1674, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 243452, "num_examples": 1630, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 384894725, "num_examples": 3100000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fr.70.jsonl.bz2": {"num_bytes": 183308200, "checksum": "9ae36c666dfb1e3a30fd9d33a2b5b2329f1ff498aef261b6a889e4767eb79b67"}}, "download_size": 187051416, "post_processing_size": null, "dataset_size": 385673443, "size_in_bytes": 572724859}, "fr.80": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fr.80", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 147905, "num_examples": 1007, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 155527, "num_examples": 997, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 231834, "num_examples": 1674, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 243452, "num_examples": 1630, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 608206389, "num_examples": 4800000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fr.70.jsonl.bz2": {"num_bytes": 183308200, "checksum": "9ae36c666dfb1e3a30fd9d33a2b5b2329f1ff498aef261b6a889e4767eb79b67"}}, "download_size": 187051416, "post_processing_size": null, "dataset_size": 608985107, "size_in_bytes": 796036523}, "fr.75": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fr.75", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 147905, "num_examples": 1007, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 155527, "num_examples": 997, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 231834, "num_examples": 1674, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 243452, "num_examples": 1630, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 1403427568, "num_examples": 11000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fr.70.jsonl.bz2": {"num_bytes": 183308200, "checksum": "9ae36c666dfb1e3a30fd9d33a2b5b2329f1ff498aef261b6a889e4767eb79b67"}}, "download_size": 187051416, "post_processing_size": null, "dataset_size": 1404206286, "size_in_bytes": 1591257702}, "fr.70": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fr.70", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 147905, "num_examples": 1007, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 155527, "num_examples": 997, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 231834, "num_examples": 1674, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 243452, "num_examples": 1630, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 1648065205, "num_examples": 13000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fr.70.jsonl.bz2": {"num_bytes": 183308200, "checksum": "9ae36c666dfb1e3a30fd9d33a2b5b2329f1ff498aef261b6a889e4767eb79b67"}}, "download_size": 187051416, "post_processing_size": null, "dataset_size": 1648843923, "size_in_bytes": 1835895339}, "fr.65": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fr.65", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 147905, "num_examples": 1007, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 155527, "num_examples": 997, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 231834, "num_examples": 1674, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 243452, "num_examples": 1630, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 2065414543, "num_examples": 16500000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fr.60.jsonl.bz2": {"num_bytes": 303614020, "checksum": "f7484213dffc639cbcf35a0eb6b564b04b7c7650d95444d5a33d6f8940304a6a"}}, "download_size": 307357236, "post_processing_size": null, "dataset_size": 2066193261, "size_in_bytes": 2373550497}, "fr.60": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "fr.60", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 147905, "num_examples": 1007, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 155527, "num_examples": 997, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 231834, "num_examples": 1674, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 243452, "num_examples": 1630, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 2695086051, "num_examples": 22000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fr.60.jsonl.bz2": {"num_bytes": 303614020, "checksum": "f7484213dffc639cbcf35a0eb6b564b04b7c7650d95444d5a33d6f8940304a6a"}}, "download_size": 307357236, "post_processing_size": null, "dataset_size": 2695864769, "size_in_bytes": 3003222005}, "ru.100": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "ru.100", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 188221, "num_examples": 1068, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 201040, "num_examples": 1020, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 320224, "num_examples": 1855, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 355681, "num_examples": 1854, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_fr.60.jsonl.bz2": {"num_bytes": 303614020, "checksum": "f7484213dffc639cbcf35a0eb6b564b04b7c7650d95444d5a33d6f8940304a6a"}}, "download_size": 307357236, "post_processing_size": null, "dataset_size": 1065166, "size_in_bytes": 308422402}, "ru.95": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "ru.95", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 188221, "num_examples": 1068, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 201040, "num_examples": 1020, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 320224, "num_examples": 1855, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 355681, "num_examples": 1854, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 21699584, "num_examples": 150000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_ru.70.jsonl.bz2": {"num_bytes": 80356597, "checksum": "14825e5dbe97976f5085ecc89c0e3c4a0b13b5de78fd94ae23b912cb569cdf32"}}, "download_size": 84099813, "post_processing_size": null, "dataset_size": 22764750, "size_in_bytes": 106864563}, "ru.90": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "ru.90", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 188221, "num_examples": 1068, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 201040, "num_examples": 1020, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 320224, "num_examples": 1855, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 355681, "num_examples": 1854, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 24537939, "num_examples": 170000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_ru.70.jsonl.bz2": {"num_bytes": 80356597, "checksum": "14825e5dbe97976f5085ecc89c0e3c4a0b13b5de78fd94ae23b912cb569cdf32"}}, "download_size": 84099813, "post_processing_size": null, "dataset_size": 25603105, "size_in_bytes": 109702918}, "ru.85": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "ru.85", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 188221, "num_examples": 1068, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 201040, "num_examples": 1020, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 320224, "num_examples": 1855, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 355681, "num_examples": 1854, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 203731331, "num_examples": 1400000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_ru.70.jsonl.bz2": {"num_bytes": 80356597, "checksum": "14825e5dbe97976f5085ecc89c0e3c4a0b13b5de78fd94ae23b912cb569cdf32"}}, "download_size": 84099813, "post_processing_size": null, "dataset_size": 204796497, "size_in_bytes": 288896310}, "ru.80": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "ru.80", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 188221, "num_examples": 1068, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 201040, "num_examples": 1020, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 320224, "num_examples": 1855, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 355681, "num_examples": 1854, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 353022994, "num_examples": 2300000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_ru.70.jsonl.bz2": {"num_bytes": 80356597, "checksum": "14825e5dbe97976f5085ecc89c0e3c4a0b13b5de78fd94ae23b912cb569cdf32"}}, "download_size": 84099813, "post_processing_size": null, "dataset_size": 354088160, "size_in_bytes": 438187973}, "ru.75": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "ru.75", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 188221, "num_examples": 1068, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 201040, "num_examples": 1020, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 320224, "num_examples": 1855, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 355681, "num_examples": 1854, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 543461259, "num_examples": 3400000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_ru.70.jsonl.bz2": {"num_bytes": 80356597, "checksum": "14825e5dbe97976f5085ecc89c0e3c4a0b13b5de78fd94ae23b912cb569cdf32"}}, "download_size": 84099813, "post_processing_size": null, "dataset_size": 544526425, "size_in_bytes": 628626238}, "ru.70": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "ru.70", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 188221, "num_examples": 1068, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 201040, "num_examples": 1020, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 320224, "num_examples": 1855, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 355681, "num_examples": 1854, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 902093009, "num_examples": 5500000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_ru.70.jsonl.bz2": {"num_bytes": 80356597, "checksum": "14825e5dbe97976f5085ecc89c0e3c4a0b13b5de78fd94ae23b912cb569cdf32"}}, "download_size": 84099813, "post_processing_size": null, "dataset_size": 903158175, "size_in_bytes": 987257988}, "ru.65": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "ru.65", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 188221, "num_examples": 1068, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 201040, "num_examples": 1020, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 320224, "num_examples": 1855, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 355681, "num_examples": 1854, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 986402768, "num_examples": 6000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_ru.60.jsonl.bz2": {"num_bytes": 203182649, "checksum": "cba1503bce783c28686a42dc352578fb23e6159bfb84c7814158aff2237351c5"}}, "download_size": 206925865, "post_processing_size": null, "dataset_size": 987467934, "size_in_bytes": 1194393799}, "ru.60": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "ru.60", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 188221, "num_examples": 1068, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 201040, "num_examples": 1020, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 320224, "num_examples": 1855, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 355681, "num_examples": 1854, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 2394129355, "num_examples": 15000000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_ru.60.jsonl.bz2": {"num_bytes": 203182649, "checksum": "cba1503bce783c28686a42dc352578fb23e6159bfb84c7814158aff2237351c5"}}, "download_size": 206925865, "post_processing_size": null, "dataset_size": 2395194521, "size_in_bytes": 2602120386}, "sv.100": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "sv.100", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122404, "num_examples": 947, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 131687, "num_examples": 984, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 241322, "num_examples": 1901, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 246545, "num_examples": 1887, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_ru.60.jsonl.bz2": {"num_bytes": 203182649, "checksum": "cba1503bce783c28686a42dc352578fb23e6159bfb84c7814158aff2237351c5"}}, "download_size": 206925865, "post_processing_size": null, "dataset_size": 741958, "size_in_bytes": 207667823}, "sv.95": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "sv.95", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122404, "num_examples": 947, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 131687, "num_examples": 984, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 241322, "num_examples": 1901, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 246545, "num_examples": 1887, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 27127233, "num_examples": 240000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_sv.70.jsonl.bz2": {"num_bytes": 23071973, "checksum": "c9ee8e84a63716c117a410cab22717719931df3498760c547a91afecdb21b996"}}, "download_size": 26815189, "post_processing_size": null, "dataset_size": 27869191, "size_in_bytes": 54684380}, "sv.90": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "sv.90", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122404, "num_examples": 947, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 131687, "num_examples": 984, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 241322, "num_examples": 1901, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 246545, "num_examples": 1887, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 67504460, "num_examples": 600000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_sv.70.jsonl.bz2": {"num_bytes": 23071973, "checksum": "c9ee8e84a63716c117a410cab22717719931df3498760c547a91afecdb21b996"}}, "download_size": 26815189, "post_processing_size": null, "dataset_size": 68246418, "size_in_bytes": 95061607}, "sv.85": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "sv.85", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122404, "num_examples": 947, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 131687, "num_examples": 984, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 241322, "num_examples": 1901, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 246545, "num_examples": 1887, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 102236134, "num_examples": 900000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_sv.70.jsonl.bz2": {"num_bytes": 23071973, "checksum": "c9ee8e84a63716c117a410cab22717719931df3498760c547a91afecdb21b996"}}, "download_size": 26815189, "post_processing_size": null, "dataset_size": 102978092, "size_in_bytes": 129793281}, "sv.80": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "sv.80", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122404, "num_examples": 947, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 131687, "num_examples": 984, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 241322, "num_examples": 1901, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 246545, "num_examples": 1887, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 140067364, "num_examples": 1200000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_sv.70.jsonl.bz2": {"num_bytes": 23071973, "checksum": "c9ee8e84a63716c117a410cab22717719931df3498760c547a91afecdb21b996"}}, "download_size": 26815189, "post_processing_size": null, "dataset_size": 140809322, "size_in_bytes": 167624511}, "sv.75": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "sv.75", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122404, "num_examples": 947, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 131687, "num_examples": 984, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 241322, "num_examples": 1901, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 246545, "num_examples": 1887, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 165122167, "num_examples": 1400000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_sv.70.jsonl.bz2": {"num_bytes": 23071973, "checksum": "c9ee8e84a63716c117a410cab22717719931df3498760c547a91afecdb21b996"}}, "download_size": 26815189, "post_processing_size": null, "dataset_size": 165864125, "size_in_bytes": 192679314}, "sv.70": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "sv.70", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122404, "num_examples": 947, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 131687, "num_examples": 984, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 241322, "num_examples": 1901, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 246545, "num_examples": 1887, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 190113575, "num_examples": 1600000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_sv.70.jsonl.bz2": {"num_bytes": 23071973, "checksum": "c9ee8e84a63716c117a410cab22717719931df3498760c547a91afecdb21b996"}}, "download_size": 26815189, "post_processing_size": null, "dataset_size": 190855533, "size_in_bytes": 217670722}, "sv.65": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "sv.65", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122404, "num_examples": 947, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 131687, "num_examples": 984, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 241322, "num_examples": 1901, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 246545, "num_examples": 1887, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 252689469, "num_examples": 2100000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_sv.60.jsonl.bz2": {"num_bytes": 60666976, "checksum": "c9751cf1f906e99e3caa553acdb7c8c7d22bdb2d9875c507f807c51885429f4f"}}, "download_size": 64410192, "post_processing_size": null, "dataset_size": 253431427, "size_in_bytes": 317841619}, "sv.60": {"description": "Opusparcus is a paraphrase corpus for six European languages: German,\nEnglish, Finnish, French, Russian, and Swedish. The paraphrases are\nextracted from the OpenSubtitles2016 corpus, which contains subtitles\nfrom movies and TV shows.\n", "citation": "@InProceedings{creutz:lrec2018,\n title = {Open Subtitles Paraphrase Corpus for Six Languages},\n author={Mathias Creutz},\n booktitle={Proceedings of the 11th edition of the Language Resources\n and Evaluation Conference (LREC 2018)},\n year={2018},\n month = {May 7-12},\n address = {Miyazaki, Japan},\n editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri\n and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti\n Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and\n H\u00e9l\u00e8ne Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis\n and Takenobu Tokunaga},\n publisher = {European Language Resources Association (ELRA)},\n isbn = {979-10-95546-00-9},\n language = {english},\n url={http://www.lrec-conf.org/proceedings/lrec2018/pdf/131.pdf}\n}\n", "homepage": "http://urn.fi/urn:nbn:fi:lb-2018021221", "license": "CC-BY-NC", "features": {"lang": {"dtype": "string", "id": null, "_type": "Value"}, "input": {"dtype": "string", "id": null, "_type": "Value"}, "target": {"dtype": "string", "id": null, "_type": "Value"}, "annot_score": {"dtype": "float32", "id": null, "_type": "Value"}, "gem_id": {"dtype": "string", "id": null, "_type": "Value"}, "references": [{"dtype": "string", "id": null, "_type": "Value"}]}, "post_processed": null, "supervised_keys": {"input": "sent1", "output": "sent2"}, "task_templates": null, "builder_name": "opusparcus", "config_name": "sv.60", "version": {"version_str": "1.0.0", "description": "", "major": 1, "minor": 0, "patch": 0}, "splits": {"test": {"name": "test", "num_bytes": 122404, "num_examples": 947, "dataset_name": "opusparcus"}, "validation": {"name": "validation", "num_bytes": 131687, "num_examples": 984, "dataset_name": "opusparcus"}, "test.full": {"name": "test.full", "num_bytes": 241322, "num_examples": 1901, "dataset_name": "opusparcus"}, "validation.full": {"name": "validation.full", "num_bytes": 246545, "num_examples": 1887, "dataset_name": "opusparcus"}, "train": {"name": "train", "num_bytes": 547241219, "num_examples": 4500000, "dataset_name": "opusparcus"}}, "download_checksums": {"validation.jsonl": {"num_bytes": 1908586, "checksum": "ca4660fa06b0fac52a37b563ea513afb0cd6bed636e3fee287645fa8711e5336"}, "test.jsonl": {"num_bytes": 1834630, "checksum": "3566d2537e3e2c40509ca373b3cd9a06fa999529bd1cd615634e35cdecd98980"}, "train_sv.60.jsonl.bz2": {"num_bytes": 60666976, "checksum": "c9751cf1f906e99e3caa553acdb7c8c7d22bdb2d9875c507f807c51885429f4f"}}, "download_size": 64410192, "post_processing_size": null, "dataset_size": 547983177, "size_in_bytes": 612393369}} |