BSC-LT/

CAESAR-TINY

	Dataset Card for CAESAR-TINY




	
	Dataset Summary

CAESAR-TINY is a synthetic code-switched dataset generated by combining monolingual samples in Catalan and Spanish. The process includes trimming silences, normalizing audio volume, and introducing random pauses. It contains 2 hours of speech data, created by concatenating audio from the Common voice 17 Benchmark split and VoxForge Spanish datasets.

To load CAESAR-TINY: from datasets import… See the full description on the dataset page: https://huggingface.co/datasets/BSC-LT/CAESAR-TINY.