AudioMNIST for SpeechBrain - Enhanced
Marcel Ackermann Sören Becker · Zenodo (CERN European Organization for Nuclear Research) · 2022
AudioMNIST dataset, adapted for SpeechBrain. The following transformations have been applied to the original data: - Trimming initial and final silence - Simple amplitude normalization: \(x_{\text{norm}} = \frac{x}{\max{|x|}}\) https://arxiv.org/abs/1807.03418