Sound-Recognition-Tutorial
Sound-Recognition-Tutorial copied to clipboard
JasonZhang156
A simple sound recognition tutorial, including data analysis, feature extraction, model building, model train and model test ...
The attribute of "lirosa.logamplitue" has been replaced by "librosa.core.amplitude_to_db" in Librosa latest release, and its argu of "ref_power" should change to "ref". D = librosa.core.amplitude_to_db(np.abs(librosa.stft(y)) ** 2, ref=np.max)
评测函数有问题
keras里面的评测的ACC本质上计算是这样的：label：[[0,1,0,0],[0,0,0,1]]，pred:[[0,0,0,0],[0,0,0,1]]准确率不是50%,而是87.5%
Metadata
Owner
Metadata
A simple sound recognition tutorial, including data analysis, feature extraction, model building, model train and model test ...