Some samples generated by WaveAI:

audio_881812_ecc85d5f3f550e078f57.wav

audio_1008669_2edeeae6de198c389ede.wav

audio_954006_50995d082f1ebdb10d64.wav

audio_943693_f0ed5c87343c642e154e.wav

audio_902439_d76921e21ee213c8aaf7.wav

audio_995260_a23526c011504d9204db.wav

audio_785895_c33c756fef474e4c20f6.wav

The model is currently in active development, so results will continue to improve. Note that there is no lyrics conditioning at this stage.

Since no dropout was used during this training, the results should be viewed with caution.