F5-TTS training dataset for Nagamese (nag). Contains 9,688 clips at 24kHz (SNR >= 20dB) from the cleaned NE-TTS dataset, formatted for F5-TTS training.

	Stats

Metric Value

Clips 9,688

Hours 14.5h

Sample rate 24kHz

SNR filter

= 20dB

Source ne-tts-nag

	Schema

Column Type Description

audio Audio 24kHz WAV audio

text string Cleaned transcript

Note

Audio is upsampled from… See the full description on the dataset page: https://huggingface.co/datasets/sulabhkatiyar/ne-tts-f5-nag.