Effective-Audio-Classification-Network-Based-on-Paired-Inverse-Pyramid-Structure-and-Dense-MLP-Block

The repetition procedure goes in this way:

• Preprocess Your UrbanSound8k Data into five features with size of (399,100) by virutal of Spafe Lib (as it is shown in the code)
• Load the Model and Processed Data into your computer memory for less training time.
• Wait until the training process finishes.

The results have been updated in the new version of arxiv, the comparison code will be released once the thesis is updated. https://arxiv.org/abs/2211.02940

Upd:2023.3.17. Upload Comparison With Other Models