Hello,

for avoiding problems, you should check that your Chinese files are encoded as utf-8. Moreover, it is recommended to apply some word segmenter to the Chinese corpora, prior to any other operation (e.g. the Standford segmenter).

from nmt-keras.

Thank you .
Does the demo-web using the model that trained from EuTrans?
Another question,I have tried use the demo-web,but when I click the translate button,there returned some php wrong:

Maybe I forget install something?

from nmt-keras.

`<?php

make sure that only one script at a time accesses the sampling server

$fp = fopen('/tmp/flock', 'w');

function file_get_contents_utf8($fn) {
$content = file_get_contents($fn);
return mb_convert_encoding($content, 'UTF-8',
mb_detect_encoding($content, 'UTF-8, ISO-8859-1', true));
}

if (flock($fp, LOCK_EX))
{
$source=$_GET['source'];
$url = 'http://localhost:8888/?source='.urlencode($source);
$out = file_get_contents($url);
echo $out;
}
else
{
echo "Server timeout! Try again later.";
}

?>`

from nmt-keras.

Did you follow the steps required for running the demo?: https://github.com/lvapeab/nmt-keras/tree/master/demo-web

If so, you must check your php configuration

from nmt-keras.

Very thank you.I am committed to using Keras to achieve Chinese and English translation.I will have a try.

from nmt-keras.