We used (1) VGGish, a deep, pretrained, Tensorflow model as an audio feature ... One has to bear in mind that the DenseNet has been trained with 1.
確定! 回上一頁