RoxasYTB/

cores_ru-ljspeech

LJSpeech dataset of Cores (ru).

327 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 ru
--input-dir ./cores_ru
--output-dir ./train_cores_ru
--dataset-format ljspeech
--single-speaker
--sample-rate 22050

python3 -m piper_train… See the full description on the dataset page: https://huggingface.co/datasets/RoxasYTB/cores_ru-ljspeech.