About the detail of your method？

[KangolHsu]

can you describe the details of your method？

[NHZlX]

I've been busy lately. I'll give you a detailed description later

[NHZlX]

@KangolHsu It's recommended to look at the new added layers.

[NHZlX]

@KangolHsu if you want know the details of the auto pruning , refer here

[KangolHsu]

@NHZlX
Thank you！！！ I kown how to deploy quantized CNN model，but i am comfuse about deploying pruned CNN model。 are we set pruned weights to 0 when deploying CNN model？

[NHZlX]

@KangolHsu This pruning cuts the paramters a in a random way, though it can be accelerated by sparse matrix multiplication, but , it's hard to speed up on embedded devices. In most cases, it's used to reduce the size of the app. Pruning and quantization are complementary to each other, so, we can apply both of them to our model.