German language model looking for helping hands (contributors)

Hey community,

@synesthesiam started to create a Google Doc Spreadsheet with about 2000 german sentences. By now I’ve started to check the sentences for three criterias:

  1. vulgar or offensive
  2. don’t make sense / make sense
  3. are something no one would actually say

I"m looking for some helping hands.

In case you would like to help out just leave a post underneath or write me a PN so I can get in touch with you.

The quicker we can verify the sentences, the sooner @synesthesiam can make a Google voice to evaluate the approach. So let’s work together and push this thing forward :slight_smile:

2 Likes

I can verify sentences, jut not good at reading out aloud, so I will be no help in actually reading stuff for training data (unless you want way to fast speech with misread words because my mouth is slower than my brain).

1 Like

I want to help with verifying the sentences :slight_smile:

2 Likes

I can also help with the sentence verification.

I could help aswell with verification :smile:
Recording is harder because i don’t have a good enough mic.

I can also help with the sentence verification.

Hi all!
Small update:

Currently there are three people working on this, all ~2500 sentences are checked, sorted, cleaned up and waiting for a second recheck.

After that we commit them to the Git.

We keep you updated :slight_smile:

4 Likes

Count me in if there’s anything to do

Hey @schnopsi,
thank you very much for wanting to contribute. @synesthesiam already used the team’s results to train the Larynx model de_thorsten which should already be available in Rhasspy’s TTS settings.

2 Likes