1 dataset found

MixSpeech is a data augmentation method for automatic speech recognition, which trains an ASR model by taking a weighted combination of two different speech features as the...