资源分类
》 资源 》算法》ko_en_NMT

ko_en_NMT

2019-09-19 | |  49 |   0 |   0

Neural Korean to English Machine Translater with Gluon

2018 .

  • Seq2Seq with Attention

  • Gluon Hybridize

  • (from https://github.com/jungyeul/korean-parallel-corpora)

  • multi gpu

  • Variable Sentence Length

  • Python 3.5

  • MXNet 1.0

  • tqdm, konlpy, stemming, gensim

  • seq2seq

  • Attention

  • bi-directional GRU encoder

TODO

  • Beam Search

How to


python main.py --embedding


CUDA_VISIBLE_DEVICES='4,5,6,7' python main.py --train --gpu-count 4 --batch-size 100


python main.py --test --init-model "models/trained_model.params" 

kor >  .
 .
eng > obama is a presid .
kor >  .
 .
eng > good morn
kor >    .
   .
eng > north korea has a nuclear weapon .
kor >

  • models/trained_model.params

    1. rmsprop(batch size : 100, lr : 0.01) : 7 epocs

    2. sgd (batch size : 100, lr : 0.01) : 33 epocs


上一篇：FacialLandmark

下一篇：DarkRank

用户评价
全部评价

热门资源

  • seetafaceJNI

    项目介绍 基于中科院seetaface2进行封装的JAVA...

  • spark-corenlp

    This package wraps Stanford CoreNLP annotators ...

  • Keras-ResNeXt

    Keras ResNeXt Implementation of ResNeXt models...

  • capsnet-with-caps...

    CapsNet with capsule-wise convolution Project ...

  • inferno-boilerplate

    This is a very basic boilerplate example for pe...