|
{"computers_xlarge": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "computers_xlarge", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 69247292, "num_examples": 54768, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1375963, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 17090688, "num_examples": 13693, "dataset_name": "products2017"}}, "download_checksums": {"computers/train_xlarge.json.gz": {"num_bytes": 21979464, "checksum": "d19055be29b211d7c7efd08413e1180c91fa4f7180c4d01572af9479ac31609a"}, "computers/valid_xlarge.json.gz": {"num_bytes": 5481832, "checksum": "ce4789d79d228fd3cbe6c4a8de4027119c7c7c9332ceb73d014cd6b79c61ed96"}, "computers/test.json.gz": {"num_bytes": 440473, "checksum": "daeee50fc827d32838365da9963239917f621e367f8944ec9c90196ee529c46b"}}, "download_size": 27901769, "post_processing_size": null, "dataset_size": 87713943, "size_in_bytes": 115615712}, "computers_large": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "computers_large", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 33627310, "num_examples": 26687, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1375963, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 8099884, "num_examples": 6672, "dataset_name": "products2017"}}, "download_checksums": {"computers/train_large.json.gz": {"num_bytes": 10749627, "checksum": "27f32be2bc7ad18c9da715c81d899391f7eaf3d5b431ab261b00acc0a608a31a"}, "computers/valid_large.json.gz": {"num_bytes": 2619716, "checksum": "524ad560d4865226149ad6c783bf9654ee1d8660c2f5aac827940b198d57d625"}, "computers/test.json.gz": {"num_bytes": 440473, "checksum": "daeee50fc827d32838365da9963239917f621e367f8944ec9c90196ee529c46b"}}, "download_size": 13809816, "post_processing_size": null, "dataset_size": 43103157, "size_in_bytes": 56912973}, "computers_medium": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "computers_medium", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 8161170, "num_examples": 6475, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1375963, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 2058805, "num_examples": 1619, "dataset_name": "products2017"}}, "download_checksums": {"computers/train_medium.json.gz": {"num_bytes": 2562779, "checksum": "4b22ca10d19b3e0061bd53e22a4bed67595c981bfff38d27f231be36ed8649d8"}, "computers/valid_medium.json.gz": {"num_bytes": 660571, "checksum": "8b11017fb8560165d7cc2c4884808ac95174847e8cd27b52cf59e711dfb1bfa1"}, "computers/test.json.gz": {"num_bytes": 440473, "checksum": "daeee50fc827d32838365da9963239917f621e367f8944ec9c90196ee529c46b"}}, "download_size": 3663823, "post_processing_size": null, "dataset_size": 11595938, "size_in_bytes": 15259761}, "computers_small": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "computers_small", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 2913705, "num_examples": 2267, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1375963, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 668057, "num_examples": 567, "dataset_name": "products2017"}}, "download_checksums": {"computers/train_small.json.gz": {"num_bytes": 913653, "checksum": "013bd9c97b0b817d299e0a8ca3583692c8c0d12de0876590b16303bed12db0f2"}, "computers/valid_small.json.gz": {"num_bytes": 211906, "checksum": "01448826a98a9b4ccf63186caf021ab650fe870dcbfdc13f3de0e1df67c81367"}, "computers/test.json.gz": {"num_bytes": 440473, "checksum": "daeee50fc827d32838365da9963239917f621e367f8944ec9c90196ee529c46b"}}, "download_size": 1566032, "post_processing_size": null, "dataset_size": 4957725, "size_in_bytes": 6523757}, "cameras_xlarge": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "cameras_xlarge", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 58930668, "num_examples": 33821, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1968970, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 14718358, "num_examples": 8456, "dataset_name": "products2017"}}, "download_checksums": {"cameras/train_xlarge.json.gz": {"num_bytes": 21291906, "checksum": "c23f4bea284e030bcc6bf67b143a0252b9a88baa1868ae4b91647d1f51684072"}, "cameras/valid_xlarge.json.gz": {"num_bytes": 5338288, "checksum": "b9c0527d4275b87e8ca2d13d93059407dfb78117b5941d0c9b349394dde33dfd"}, "cameras/test.json.gz": {"num_bytes": 662236, "checksum": "bce167650dbf6c993e848baa9e2759ec91f4e0147c7032cec938eb0f6c777f2c"}}, "download_size": 27292430, "post_processing_size": null, "dataset_size": 75617996, "size_in_bytes": 102910426}, "cameras_large": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "cameras_large", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 26874867, "num_examples": 16028, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1968970, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 6588301, "num_examples": 4008, "dataset_name": "products2017"}}, "download_checksums": {"cameras/train_large.json.gz": {"num_bytes": 9702430, "checksum": "4a39dbf9f3c5e62a4ab330e5cabe2e3ffa8268765a9b56ba6e566bb5021e1077"}, "cameras/valid_large.json.gz": {"num_bytes": 2364699, "checksum": "84236bb649899400ee816fed8c1dda06c4bea78f02f8a1c08032fdef51613ba4"}, "cameras/test.json.gz": {"num_bytes": 662236, "checksum": "bce167650dbf6c993e848baa9e2759ec91f4e0147c7032cec938eb0f6c777f2c"}}, "download_size": 12729365, "post_processing_size": null, "dataset_size": 35432138, "size_in_bytes": 48161503}, "cameras_medium": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "cameras_medium", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 6985298, "num_examples": 4204, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1968970, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 1724070, "num_examples": 1051, "dataset_name": "products2017"}}, "download_checksums": {"cameras/train_medium.json.gz": {"num_bytes": 2508508, "checksum": "d0d597eb300adb2de5b860e9773bd81e23878e80ab612cf3ca9257b11ae76ce9"}, "cameras/valid_medium.json.gz": {"num_bytes": 614504, "checksum": "0bc9ffdbad348e1966bd624e3725bab54f1309452f075a7f65d973f35bed186f"}, "cameras/test.json.gz": {"num_bytes": 662236, "checksum": "bce167650dbf6c993e848baa9e2759ec91f4e0147c7032cec938eb0f6c777f2c"}}, "download_size": 3785248, "post_processing_size": null, "dataset_size": 10678338, "size_in_bytes": 14463586}, "cameras_small": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "cameras_small", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 2486310, "num_examples": 1508, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1968970, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 672835, "num_examples": 378, "dataset_name": "products2017"}}, "download_checksums": {"cameras/train_small.json.gz": {"num_bytes": 896323, "checksum": "90eafd129c495d4120b3bdc4e0004d7e0b075669b299f434f89addf25468926f"}, "cameras/valid_small.json.gz": {"num_bytes": 240242, "checksum": "a8ad16d21e69067b593a48a8fe55bc77f5e914b98520d531474d10c7877a9dec"}, "cameras/test.json.gz": {"num_bytes": 662236, "checksum": "bce167650dbf6c993e848baa9e2759ec91f4e0147c7032cec938eb0f6c777f2c"}}, "download_size": 1798801, "post_processing_size": null, "dataset_size": 5128115, "size_in_bytes": 6926916}, "watches_xlarge": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "watches_xlarge", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 84464343, "num_examples": 49255, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1836593, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 20901713, "num_examples": 12314, "dataset_name": "products2017"}}, "download_checksums": {"watches/train_xlarge.json.gz": {"num_bytes": 22312475, "checksum": "265a1a8729a69ffcb20564ffe50bae14eb86dcce8f96ad36d1fb74ab084275a5"}, "watches/valid_xlarge.json.gz": {"num_bytes": 5463237, "checksum": "8c8dc012317e2b1df8ad01e4b3adecbc07ab4387e0b2259b9d55c86c7cb089df"}, "watches/test.json.gz": {"num_bytes": 514961, "checksum": "d632c71bf50f7d127068e496acd8755247157b016a098d6b79a8c787c5b13ca4"}}, "download_size": 28290673, "post_processing_size": null, "dataset_size": 107202649, "size_in_bytes": 135493322}, "watches_large": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "watches_large", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 35867387, "num_examples": 21621, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1836593, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 8969150, "num_examples": 5406, "dataset_name": "products2017"}}, "download_checksums": {"watches/train_large.json.gz": {"num_bytes": 10263745, "checksum": "e24dcd1a732f0dee0fa3fd81024f11d1c8a0ece1738bc26efacb52b697e261e1"}, "watches/valid_large.json.gz": {"num_bytes": 2563533, "checksum": "7339ef6ea2f34e289db98609b86f2ee327d7105b3f19c3919ec0ddb7358ed1ce"}, "watches/test.json.gz": {"num_bytes": 514961, "checksum": "d632c71bf50f7d127068e496acd8755247157b016a098d6b79a8c787c5b13ca4"}}, "download_size": 13342239, "post_processing_size": null, "dataset_size": 46673130, "size_in_bytes": 60015369}, "watches_medium": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "watches_medium", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 8467978, "num_examples": 5130, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1836593, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 2103680, "num_examples": 1283, "dataset_name": "products2017"}}, "download_checksums": {"watches/train_medium.json.gz": {"num_bytes": 2526427, "checksum": "0f538119833efb1c8b565a50b1defe4aace04a20c278c10cacab6361ab091bcc"}, "watches/valid_medium.json.gz": {"num_bytes": 623683, "checksum": "a83688e1074dda7233cb4820cf82db04dd89015607c5e0003513de11e7de4b30"}, "watches/test.json.gz": {"num_bytes": 514961, "checksum": "d632c71bf50f7d127068e496acd8755247157b016a098d6b79a8c787c5b13ca4"}}, "download_size": 3665071, "post_processing_size": null, "dataset_size": 12408251, "size_in_bytes": 16073322}, "watches_small": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "watches_small", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 2901578, "num_examples": 1804, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1836593, "num_examples": 1100, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 730053, "num_examples": 451, "dataset_name": "products2017"}}, "download_checksums": {"watches/train_small.json.gz": {"num_bytes": 867101, "checksum": "7f36b2a66bb9c574adb66a047783cf20721aff21a9635610eeefc5afff3cbc8f"}, "watches/valid_small.json.gz": {"num_bytes": 213527, "checksum": "01eb5dad8a9db595fcf24cad482fbb67114eb6361c5aee42820c1ad42009fc43"}, "watches/test.json.gz": {"num_bytes": 514961, "checksum": "d632c71bf50f7d127068e496acd8755247157b016a098d6b79a8c787c5b13ca4"}}, "download_size": 1595589, "post_processing_size": null, "dataset_size": 5468224, "size_in_bytes": 7063813}, "shoes_xlarge": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "shoes_xlarge", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 45819068, "num_examples": 33943, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1590242, "num_examples": 1099, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 11559195, "num_examples": 8486, "dataset_name": "products2017"}}, "download_checksums": {"shoes/train_xlarge.json.gz": {"num_bytes": 16435876, "checksum": "1718653a3b4a8dcd47b5c1f55d7fae4e3fcf3c04af13e4446cf49a4b48f7f504"}, "shoes/valid_xlarge.json.gz": {"num_bytes": 4181524, "checksum": "461e37861da72d786ce11da3b78033fe0785969bf41df0b49ef27d6e7e5f0f27"}, "shoes/test.json.gz": {"num_bytes": 470891, "checksum": "cb5fb369bea226e8abc6d0fefcbbb74f4d4fb262774c6a9ecc3686d42c640683"}}, "download_size": 21088291, "post_processing_size": null, "dataset_size": 58968505, "size_in_bytes": 80056796}, "shoes_large": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "shoes_large", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 24288014, "num_examples": 18391, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1590242, "num_examples": 1099, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 6064641, "num_examples": 4598, "dataset_name": "products2017"}}, "download_checksums": {"shoes/train_large.json.gz": {"num_bytes": 8745243, "checksum": "fdf8c3b0f9516e7d7a24c41f6ed8fe8926b04f42f36969d148d68912acac5472"}, "shoes/valid_large.json.gz": {"num_bytes": 2160668, "checksum": "03ce9ab66ccc8569917c1ba56f627cca43780746d26ed5fd2afba9f24f155ced"}, "shoes/test.json.gz": {"num_bytes": 470891, "checksum": "cb5fb369bea226e8abc6d0fefcbbb74f4d4fb262774c6a9ecc3686d42c640683"}}, "download_size": 11376802, "post_processing_size": null, "dataset_size": 31942897, "size_in_bytes": 43319699}, "shoes_medium": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "shoes_medium", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 5876799, "num_examples": 4644, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1590242, "num_examples": 1099, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 1524284, "num_examples": 1161, "dataset_name": "products2017"}}, "download_checksums": {"shoes/train_medium.json.gz": {"num_bytes": 2123481, "checksum": "17d36cca4b6bbe6e25494245b14ddbe9323b9d4858e615ae1fa5670e02fb1c39"}, "shoes/valid_medium.json.gz": {"num_bytes": 545330, "checksum": "3b1f33af9314ea3826265af0cc4e26290da6ab8f194fd8f81d8764ae84fd9bc3"}, "shoes/test.json.gz": {"num_bytes": 470891, "checksum": "cb5fb369bea226e8abc6d0fefcbbb74f4d4fb262774c6a9ecc3686d42c640683"}}, "download_size": 3139702, "post_processing_size": null, "dataset_size": 8991325, "size_in_bytes": 12131027}, "shoes_small": {"description": "Many e-shops have started to mark-up product data within their HTML pages using the schema.org vocabulary. The Web Data Commons project regularly extracts such data from the Common Crawl, a large public web crawl. The Web Data Commons Training and Test Sets for Large-Scale Product Matching contain product offers from different e-shops in the form of binary product pairs (with corresponding label \"match\" or \"no match\")\n\nIn order to support the evaluation of machine learning-based matching methods, the data is split into training, validation and test set. We provide training and validation sets in four different sizes for four product categories. The labels of the test sets were manually checked while those of the training sets were derived using shared product identifiers from the Web via weak supervision.\n\nThe data stems from the WDC Product Data Corpus for Large-Scale Product Matching - Version 2.0 which consists of 26 million product offers originating from 79 thousand websites.\n", "citation": "@inproceedings{primpeli2019wdc,\n title={The WDC training dataset and gold standard for large-scale product matching},\n author={Primpeli, Anna and Peeters, Ralph and Bizer, Christian},\n booktitle={Companion Proceedings of The 2019 World Wide Web Conference},\n pages={381--386},\n year={2019}\n}\n", "homepage": "http://webdatacommons.org/largescaleproductcorpus/v2/index.html", "license": "", "features": {"pair_id": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"dtype": "int32", "id": null, "_type": "Value"}, "id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "category_left": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_left": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_left": {"dtype": "string", "id": null, "_type": "Value"}, "title_left": {"dtype": "string", "id": null, "_type": "Value"}, "description_left": {"dtype": "string", "id": null, "_type": "Value"}, "price_left": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_left": {"dtype": "string", "id": null, "_type": "Value"}, "id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "category_right": {"dtype": "string", "id": null, "_type": "Value"}, "cluster_id_right": {"dtype": "int32", "id": null, "_type": "Value"}, "brand_right": {"dtype": "string", "id": null, "_type": "Value"}, "title_right": {"dtype": "string", "id": null, "_type": "Value"}, "description_right": {"dtype": "string", "id": null, "_type": "Value"}, "price_right": {"dtype": "string", "id": null, "_type": "Value"}, "specTableContent_right": {"dtype": "string", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "products2017", "config_name": "shoes_small", "version": {"version_str": "2.1.0", "description": null, "major": 2, "minor": 1, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 2089371, "num_examples": 1650, "dataset_name": "products2017"}, "test": {"name": "test", "num_bytes": 1590242, "num_examples": 1099, "dataset_name": "products2017"}, "validation": {"name": "validation", "num_bytes": 531266, "num_examples": 413, "dataset_name": "products2017"}}, "download_checksums": {"shoes/train_small.json.gz": {"num_bytes": 757540, "checksum": "cd26d5e1bae5efbac8134eb784a0b2c7d2b6691195d5b7b7eeb304683e6c557d"}, "shoes/valid_small.json.gz": {"num_bytes": 194196, "checksum": "8d76d3c97888e892abbce436cc14334c862a84adc8929fd073089229c6f214b0"}, "shoes/test.json.gz": {"num_bytes": 470891, "checksum": "cb5fb369bea226e8abc6d0fefcbbb74f4d4fb262774c6a9ecc3686d42c640683"}}, "download_size": 1422627, "post_processing_size": null, "dataset_size": 4210879, "size_in_bytes": 5633506}} |
