RoxasYTB/

oracle_turret_es-ljspeech

LJSpeech dataset of Oracle Turret (es).

23 pairs 44.1kHz mono 16-bit PCM WAV (original wiki quality)

	Piper TTS Training (High Quality on T4 GPU)




	


	Preprocessing (downsample to 22.05kHz)

python3 -m piper_train.preprocess
--language es
--input-dir ./oracle_turret_es
--output-dir ./train_oracle_turret_es
--dataset-format ljspeech
--single-speaker
--sample-rate 22050