langrank:选择用于跨语言学习的传输语言的程序-源码

上传者: 42116921 | 上传时间: 2021-03-18 17:08:55 | 文件大小: 3.74MB | 文件类型: ZIP
郎朗克 通过 @ LangRank是用于为跨语言迁移学习选择传输语言的程序,我们在ACL 2019上的主题中对此进行了介绍。跨语言传输,其中使用高资源传输语言来提高低资源任务的准确性语言现在是提高自然资源处理(NLP)在低资源语言上的性能的宝贵工具。但是,给定特定的任务语言,尚不清楚要从哪种语言进行转换,标准策略是根据特殊标准(通常是实验人员的直觉)选择语言。由于大量功能有助于跨语言传输(包括系统发育相似性,类型学特性,词汇重叠或可用数据的大小)的成功,因此即使是最开明的实验者也很少考虑所有这些因素来完成手头的特定任务。 LangRank是一个程序,用于解决以下任务:自动选择最佳传输语言,将其视为排名问题,并构建考虑上述功能的模型以执行此预测。例如,假设您有一个机器翻译(MT)任务,并且想知道您应该使用哪种语言/数据集来构建资源少的语言Azerbaijani的系统。在这种情况下,您将准

文件下载

资源详情

[{"title":"( 28 个子文件 3.74MB ) langrank:选择用于跨语言学习的传输语言的程序-源码","children":[{"title":"langrank-master","children":[{"title":"figures","children":[{"title":"overview.png <span style='color:#111;'> 131.40KB </span>","children":null,"spread":false},{"title":"results.png <span style='color:#111;'> 127.10KB </span>","children":null,"spread":false}],"spread":true},{"title":"langrank.py <span style='color:#111;'> 14.01KB </span>","children":null,"spread":false},{"title":"requirements.txt <span style='color:#111;'> 9B </span>","children":null,"spread":false},{"title":"lang2vec.py <span style='color:#111;'> 0B </span>","children":null,"spread":false},{"title":"index_ted_datasets.py <span style='color:#111;'> 3.31KB </span>","children":null,"spread":false},{"title":"README.md <span style='color:#111;'> 6.83KB </span>","children":null,"spread":false},{"title":"index_pos_datasets.py <span style='color:#111;'> 4.58KB </span>","children":null,"spread":false},{"title":"index_el_datasets.py <span style='color:#111;'> 5.36KB </span>","children":null,"spread":false},{"title":"sample-data","children":[{"title":"ted-train.orig.aze <span style='color:#111;'> 580.67KB </span>","children":null,"spread":false},{"title":"ell.tok <span style='color:#111;'> 919.65KB </span>","children":null,"spread":false},{"title":"ted-train.orig.spm8000.fin <span style='color:#111;'> 3.09MB </span>","children":null,"spread":false},{"title":"ted-train.orig.fin <span style='color:#111;'> 2.10MB </span>","children":null,"spread":false},{"title":"ted-train.orig.spm8000.aze <span style='color:#111;'> 834.58KB </span>","children":null,"spread":false},{"title":"ell.tok.bpe <span style='color:#111;'> 949.97KB </span>","children":null,"spread":false},{"title":"ted-train.orig.ben <span style='color:#111;'> 1.14MB </span>","children":null,"spread":false},{"title":"ted-train.orig.spm8000.ben <span style='color:#111;'> 1.40MB </span>","children":null,"spread":false}],"spread":true},{"title":"tests","children":[{"title":"test_train_file.py <span style='color:#111;'> 791B </span>","children":null,"spread":false}],"spread":true},{"title":".gitignore <span style='color:#111;'> 18B </span>","children":null,"spread":false},{"title":"langrank_predict.py <span style='color:#111;'> 1.59KB </span>","children":null,"spread":false},{"title":"pretrained","children":[{"title":"EL","children":[{"title":"lgbm_model_el_all.txt <span style='color:#111;'> 134.79KB </span>","children":null,"spread":false}],"spread":false},{"title":"POS","children":[{"title":"lgbm_model_pos_all.txt <span style='color:#111;'> 134.51KB </span>","children":null,"spread":false}],"spread":false},{"title":"MT","children":[{"title":"lgbm_model_mt_all.txt <span style='color:#111;'> 135.54KB </span>","children":null,"spread":false},{"title":"lgbm_model_mt_aze.txt <span style='color:#111;'> 5.52KB </span>","children":null,"spread":false},{"title":"lgbm_model_mt_fin.txt <span style='color:#111;'> 10.85KB </span>","children":null,"spread":false},{"title":"lgbm_model_mt_ben.txt <span style='color:#111;'> 41.23KB </span>","children":null,"spread":false}],"spread":false},{"title":"DEP","children":[{"title":"lgbm_model_dep_all.txt <span style='color:#111;'> 134.60KB </span>","children":null,"spread":false}],"spread":false}],"spread":true},{"title":"index_parsing_datasets.py <span style='color:#111;'> 4.62KB </span>","children":null,"spread":false}],"spread":false}],"spread":true}]

评论信息

免责申明

【只为小站】的资源来自网友分享,仅供学习研究,请务必在下载后24小时内给予删除,不得用于其他任何用途,否则后果自负。基于互联网的特殊性,【只为小站】 无法对用户传输的作品、信息、内容的权属或合法性、合规性、真实性、科学性、完整权、有效性等进行实质审查;无论 【只为小站】 经营者是否已进行审查,用户均应自行承担因其传输的作品、信息、内容而可能或已经产生的侵权或权属纠纷等法律责任。
本站所有资源不代表本站的观点或立场,基于网友分享,根据中国法律《信息网络传播权保护条例》第二十二条之规定,若资源存在侵权或相关问题请联系本站客服人员,zhiweidada#qq.com,请把#换成@,本站将给予最大的支持与配合,做到及时反馈和处理。关于更多版权及免责申明参见 版权及免责申明