RoxasYTB/

defective_turret_en-ljspeech

LJSpeech dataset of Defective Turret (en).

190 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 en-us
--input-dir ./defective_turret_en
--output-dir ./train_defective_turret_en
--dataset-format ljspeech
--single-speaker
--sample-rate 22050… See the full description on the dataset page: https://huggingface.co/datasets/RoxasYTB/defective_turret_en-ljspeech.