Michael Hansen

I don't remember the icon ever showing the dictionary (did it in 2.4?) Which language are you using this for? For English, "titi" is actually in the dictionary and the...

The pickle format was chosen for serialization speed/size and because it's baked into NetworkX. Along with gzip, I saw a pretty big difference in file size and import/export speed between...

This won't be a lot of work, luckily. Every Hermes service uses the same shared library to parse command-line arguments and set up its MQTT connection :)

@koenvervloesem, do you know of a good tutorial to set up TLS in mosquitto?

The words page definitely needs some TLC. I can reformat it to make it more obvious how to add guessed words (an "Add" button appears, but is not very visible)....

The English examples aren't intentional; I have [another program](https://github.com/synesthesiam/voice2json-profiles/blob/master/bin/print_phoneme_examples.py) that generates the phoneme examples by looking at a [list of frequent German words](https://github.com/synesthesiam/de_kaldi-zamia/blob/master/frequent_words.txt) and picks the shortest words that more...

Yes, this is definitely possible. My approach would be to train a Kaldi ASR model using IPA phonemes. A challenge with Arabic looks to be the writing system. An extra...

Hi @B0ndo2, thank you for the offer and the link. It looks like the limiting factor for me is going to be audio data. I have close to 90 hours...

The datasets are pretty simple: just audio files (WAV, MP3, etc.) with spoken sentences and the text of what was said (usually in a text or CSV file). The hard...

> once I have them translated how do I commit them? You can post them here or e-mail them to me, and I can add them to the `rhasspy-profile` project....