November community campaign

[nukeador (Rubén Martín [❌ taking a break from Mozilla])]

Good Monday everyone,

This month we followed-up with our community campaigns, and as we did in October, we also run efforts through snippets and social media in the languages we had at least 1M sentences.

This time the efforts were run in: English, German, French, Spanish and Italian.

Again, Common Voice November community campaign followed the lead of the previous success last month. Last week we saw increases in the number of weekly recorded and validated clips from 1,5x to 36x from our new campaign participant, Italian!

Language	Recorded	Validated	Increase from prev. week
English	47hrs	30hrs	8x
Italian	36hrs	16hrs	36x
Spanish	18hrs	11hrs	9x
French	16hrs	9hrs	3x
German	15hrs	8hrs	1,5x

We also got 57K visitors, that’s 6 times more than the previous week.

Thanks everyone who participated and supported this effort!

Cheers.

[stergro (Stefan Grotz)]

That’s great. Any idea why German has only an increase of 1,5 this time?

Also congratulations for 1k hours of validated voices and 50k donators in the English dataset! I made this screenshot on Saturday:

[nukeador (Rubén Martín [❌ taking a break from Mozilla])]

That’s great. Any idea why German has only an increase of 1,5 this time?

Apparently Firefox in German had a lot of marketing snippets also last week and that’s why we reached way less people.

[nukeador (Rubén Martín [❌ taking a break from Mozilla])]

As a follow-up, we plan to run again a snippets campaign next week (Dec 9th - Dec 15th) in the same languages.

[EwoutH (Ewout ter Hoeven)]

@nukeador How far is Dutch from having 1M sentences?

[nukeador (Rubén Martín [❌ taking a break from Mozilla])]

Dutch currently has around 5K sentences. My recommendation on the easiest way to get 1M (or close) sentences is to get technical people to help with the wikipedia extractor in Dutch

[Technical feedback needed] Wikipedia extractor script beta

Hello everyone, As we announced a few months ago, as an effort to get sentences from large sources of data, we found a legal way to extract some wikipedia sentences for our corpus under Public Domain, so we can have enough volume to accommodate the minimum 2000hrs of voice we want to collect to create a basic model for speech recognition. Today I want to share the script we have been developing, so the more technical people can take an early look, play with it and help us improve so it can be …

This is how we managed to get Spanish, German, French, Italian or Catalan with that amount of sentences. Please, let me know if you have any questions.

Cheers.