自然語言處理N天-使用Pytorch實現Transformer
https://www.jianshu.com/p/e05ec4bdc60b
https://www.jianshu.com/p/4e94690ba8e3
https://www.jianshu.com/p/2eb21de7fd5f
PaddlePaddle實戰 | 千行代碼搞定Transformer
Github 上 Star 過千的 NLP 相關項目
https://opennmt.net/OpenNMT-py/main.html#installation
AllenNLP 使用教程
其中的一篇:
https://www.manning.com/books/real-world-natural-language-processing
博客地址:http://blog.csdn.net/wangxinginnlp/article/details/52944432
工具名稱:T2T: Tensor2Tensor Transformers
地址:https://github.com/tensorflow/tensor2tensor
語言:Python/Tensorflow
簡介:★★★★★ 五顆星
https://research.googleblog.com/2017/06/accelerating-deep-learning-research.html
工具名稱:dl4mt
地址:https://github.com/nyu-dl/dl4mt-tutorial/tree/master/session2
語言:Python/Theano
簡介:
Attention-based encoder-decoder model for machine translation.
New York University Kyunghyun Cho博士組開發。
工具名稱:blocks
地址:https://github.com/mila-udem/blocks
語言:Python/Theano
簡介:
Blocks is a framework that helps you build neural network models on top of Theano.
Université de Montréal LISA Lab(實驗室主任Yoshua Bengio,實驗室現在更名為MILA Lab,主頁:https://mila.umontreal.ca/en/)開發,是之前GroundHog(https://github.com/lisa-groundhog/GroundHog)的升級替代版。
工具名稱:EUREKA-MangoNMT
地址:https://github.com/jiajunzhangnlp/EUREKA-MangoNMT
語言:C++
簡介:A C++ toolkit for neural machine translation for CPU.
中科院自動化所語音語言技術研究組張家俊博士(http://www.nlpr.ia.ac.cn/cip/jjzhang.htm)開發。
工具名稱:Nematus
地址:https://github.com/EdinburghNLP/nematus
語言:Python/Theano
簡介:愛丁堡大學發布的NMT工具
工具名稱:AmuNMT
地址:https://github.com/emjotde/amunmt
語言:C++
簡介:
A C++ inference engine for Neural Machine Translation (NMT) models trained with Theano-based scripts from Nematus (https://github.com/rsennrich/nematus) or DL4MT (https://github.com/nyu-dl/dl4mt-tutorial).
Moses Machine Translation CIC公司Hieu Hoang博士(http://statmt.org/~s0565741/)等人開發。
工具名稱:Zoph_RNN
地址:https://github.com/isi-nlp/Zoph_RNN
語言:C++
簡介:
A C++/CUDA toolkit for training sequence and sequence-to-sequence models across multiple GPUs.
USC Information Sciences Institute開發。
工具名稱:sequence-to-sequence mdoels in tensorflow
地址:https://www.tensorflow.org/versions/r0.11/tutorials/seq2seq/index.html
語言:TensorFlow/Python
簡介:Sequence-to-Sequence Models
工具名稱:nmt_stanford_nlp
地址:http://nlp.stanford.edu/projects/nmt/
語言:Matlab
簡介:
Neural machine translation (NMT) at Stanford NLP group.
工具名稱:OpenNMT
地址:http://opennmt.net/
語言:Lua/Torch
簡介:
OpenNMT was originally developed by Yoon Kim and harvardnlp.
工具名稱:lamtram
地址:https://github.com/neubig/lamtram
語言:C++/DyNet
簡介:
lamtram: A toolkit for language and translation modeling using neural networks.
CMU Graham Neubig博士組開發。
工具名稱:Neural Monkey
地址:https://github.com/ufal/neuralmonkey
語言:TensorFlow/Python
簡介:The Neural Monkey package provides a higher level abstraction for sequential neural network models, most prominently in Natural Language Processing (NLP). It is built on TensorFlow. It can be used for fast prototyping of sequential models in NLP which can be used e.g. for neural machine translation or sentence classification.
Institute of Formal and Applied Linguistics at Charles University 開發。
(WMT中NEURAL MT TRAINING TASK用的就是Neural Monkey 見:http://www.statmt.org/wmt17/)
工具名稱:Neural Machine Translation (seq2seq) Tutorial
地址:https://github.com/tensorflow/nmt
語言:python/Tensorflow
簡介:
Google Brain的Thang Luong博士等人出品
如果對上述工具感興趣,可以使用WMT16的雙語語料跑着玩玩,語料地址 http://www.statmt.org/wmt16/translation-task.html。