Multi30k Dataset
收藏资源简介:
Multi30k数据集是一个多语言的英语-德语图像描述数据集,包含训练、验证和测试集,支持多种语言如英语、德语、法语和捷克语。数据集提供了详细的统计信息,如句子数量、单词数量和平均单词数每句。此外,还提供了视觉特征和原始图像的下载链接。
The Multi30k dataset is a multilingual English-German image description dataset, encompassing training, validation, and test sets, and supports multiple languages such as English, German, French, and Czech. The dataset provides detailed statistical information, including the number of sentences, the number of words, and the average number of words per sentence. Additionally, it offers download links for visual features and original images.
数据集概述
名称: Multi30k Data Repository
数据结构:
- Task 1:
- Raw files: 位于
data/task1/raw - Tokenized files: 位于
data/task1/tok,使用脚本scripts/task1-tokenize.sh进行预处理
- Raw files: 位于
数据集统计:
- 训练集:
- 英语 (en): 29000 句,377534 词,平均每句13.0词
- 德语 (de): 29000 句,360706 词,平均每句12.4词
- 法语 (fr): 29000 句,409845 词,平均每句14.1词
- 捷克语 (cs): 29000 句,297212 词,平均每句10.2词
- 验证集:
- 英语 (en): 1014 句,13308 词,平均每句13.1词
- 德语 (de): 1014 句,12828 词,平均每句12.7词
- 法语 (fr): 1014 句,14381 词,平均每句14.2词
- 捷克语 (cs): 1014 句,10342 词,平均每句10.2词
- 测试集:
- 2016 Flickr:
- 英语 (en): 1000 句,12968 词,平均每句13.0词
- 德语 (de): 1000 句,12103 词,平均每句12.1词
- 法语 (fr): 1000 句,13988 词,平均每句14.0词
- 捷克语 (cs): 1000 句,10497 词,平均每句10.5词
- 2017 Flickr:
- 英语 (en): 1000 句,11376 词,平均每句11.4词
- 德语 (de): 1000 句,10758 词,平均每句10.8词
- 法语 (fr): 1000 句,12596 词,平均每句12.6词
- 2017 MSCOCO:
- 英语 (en): 461 句,5239 词,平均每句11.4词
- 德语 (de): 461 句,5158 词,平均每句11.2词
- 法语 (fr): 461 句,5710 词,平均每句12.4词
- 2016 Flickr:
评估:
- 可使用 Codalab 竞赛评估模型在2018测试集上的表现。
视觉特征:
- 预提取的视觉特征可从 Google Drive 下载。
- 原始图像可通过 此链接 请求。
引用:
-
英语和德语数据:
@InProceedings{W16-3210, author = "Elliott, Desmond and Frank, Stella and Simaan, Khalil and Specia, Lucia", title = "Multi30K: Multilingual English-German Image Descriptions", booktitle = "Proceedings of the 5th Workshop on Vision and Language", year = "2016", publisher = "Association for Computational Linguistics", pages = "70--74", location = "Berlin, Germany", doi = "10.18653/v1/W16-3210", url = "http://www.aclweb.org/anthology/W16-3210" }
-
法语数据、模糊COCO评估数据和2017测试数据:
@InProceedings{elliott-EtAl:2017:WMT, author = {Elliott, Desmond and Frank, Stella and Barrault, Lo"{i}c and Bougares, Fethi and Specia, Lucia}, title = {Findings of the Second Shared Task on Multimodal Machine Translation and Multilingual Image Description}, booktitle = {Proceedings of the Second Conference on Machine Translation, Volume 2: Shared Task Papers}, month = {September}, year = {2017}, address = {Copenhagen, Denmark}, publisher = {Association for Computational Linguistics}, pages = {215--233}, url = {http://www.aclweb.org/anthology/W17-4718} }
-
捷克语数据:
@inproceedings{barrault2018findings, title={Findings of the Third Shared Task on Multimodal Machine Translation}, author={Barrault, Lo{"i}c and Bougares, Fethi and Specia, Lucia and Lala, Chiraag and Elliott, Desmond and Frank, Stella}, booktitle={Proceedings of the Third Conference on Machine Translation: Shared Task Papers}, pages={304--323}, year={2018} }




