What does the translation workflow look like? about ml4a.github.io

[genekogan]

would be great to have a German translation!
haven't seen deepl yet... how does it compare to google translate?
i think for technical subjects, ML translations usually come out sounding a bit funny still. but certainly i think any translator is free to use any strategy that works for them. right now the workflow is really simple -- just copy the english chapter into a new folder (we can make _chapters_de) and then just translate the english into german.
there are no guidelines really, translator is free to translate very strict to the original text or can get more creative and modify parts of it. especially where things don't easily translate (for example, there's a lot of english/american idioms that don't always work well translated).
let me know if you'd like to give the neural nets chapter a try (or any of the others)!

[HerrBertling]

The rather advertise-y quote underneath the translator pretty much sums up my experience so far:

Tech giants Google, Microsoft and Facebook are all applying the lessons of machine learning to translation, but a small company called DeepL has outdone them all and raised the bar for the field.
Its translation tool is just as quick as the outsized competition, but more accurate and nuanced than any we’ve tried.

And if it outputs rubbish, I guess I'll be able to adjust that. I'll give the neural nets chapter a try, then :)