The Esperanto dataset already has more than 3x the recorded audio than in the last release (90h validated vs now 309h). I’m currently working on STT for Esperanto and a interim release with the ca. 220 hours additional audio would help a lot.
1 Like