Oh and for anyone who doesn't know yet - there is this website https://forvo.com/ which has a lot of audio recordings from native speakers. You can search for a single word or a full phrase. It really helped me with Korean and German when I had doubts:). Source: 6 months ago

Another useful site for hearing pronunciations is Forvo: https://forvo.com/ Those are user contributed pronunciations, so there was an effort to say the word clearly. Although Youglish might be more authentic in a sense, I prefer hearing a word enunciated precisely if I want to learn the pronunciation. - Source: Hacker News / 8 months ago

Forvo to hear isolated recordings of words, YouGlish to hear them in context. Source: 12 months ago

Another possible resource is a site called forvo in which people pronounce words and sentences in their own languages. Very useful tool to learn pronunciations of new words but please bear in mind that sometimes they can be unrealistic if they are exaggerated and/or out of context. Source: 12 months ago

For individual words and phrases, go to http://forvo.com where you can hear native speakers in dozens of languages and even submit new words, names, or phrases. Source: 12 months ago

This is very cool, looking forward to it! I've been doing the same thing with Spanish Wikipedia articles for a while, using a few lines of Bash + Regex. I was using Apertium for it. https://apertium.org/ It's definitely worse than most ML-based solutions, but it works reliably and fast; you can run it entirely offline. With Spanish translations, the main problem I was facing is lack of vocabulary, so I created - Source: Hacker News / 10 months ago

I used to keep track of the state of machine translation some years back. I think the way you measure the success of an automated translation is edit distance, i.e. How many manual edits you need to make to a translated text before you reach some acceptable state. I suppose it's somewhat subjective, but it is possible to construct a benchmark and allow for multiple correct results. The best resources I knew back... - Source: Hacker News / about 2 years ago

Apertium is one of them. We make open-source rule-based machine translation systems, and our core tools are in C++. A few of our proposed ideas involve modifying those C++ tools with new features or improvements to existing features. Source: over 3 years ago