OpenNMT Forum
Log In
Support
Why do Source tokens/s 1 in every set of iterations?
(4)
Validation Data
(16)
Unexpected translation of tagged text
(5)
Limiting number of CPU cores / CPU usage in training
(3)
Replacement untranslated words doesn't work
(2)
How do word_features work?
(2)
If there is any way to keep placeholders as same as source when call NMT
(2)
Input vector as input and branch encoder
(9)
OutOfMemory while loading a GPU model?
(3)
Add more XML tags to tokenized corpus
(2)
How do I use Crayon?
(2)
Translation of complex numbers
(5)
What is the Kaldi input format? Where can I find an example or definition?
(4)
Why Variable rewrap / unwrap in memoryEfficientLoss()?
(1)
How to reset the learning rate?
(4)
Bpe for chinese characters?
(6)
Strange behaviour of rest_translation_server.lua
(4)
Torch/nn update breaks loading OpenNMT models trained with earlier version(?)
(7)
Out of memory when translating?
(10)
ZH->EN uppercased the second word of the sentence?
(7)
Is there a way to tell not to check part of sentence
(2)
The error when continuing a training
(2)
The warning of nccl
(3)
Sentence length & translation quality
(13)
The WMT14 English-French result on the Opennmt-py
(4)
Output from CTranslate and OpenNMT translate differ
(3)
CJK tokenizer question
(1)
How to use new feature:target_subdict
(3)
Could OpenNMT run on the MAC OXS?
(9)
Incremental training - size of new training data and vocabulary updating
(3)
← previous page
next page →
Website
Documentation
Gitter
GitHub