Model compression is an important topic in deep learning research. It can be mainly divided into two directions: model pruning and model quantization. However. both methods will more or less affect the original accuracy of the model. https://allfixelectricales.shop/product-category/fridge-door-shelf-rail/