A way to shrink a model by removing extra links or neurons.
Network pruning is like trimming a hedge before it swallows the mailbox. Snip the extra twigs, and the hedge still looks fine.
It removes weak links or neurons for Model Compression and phone-sized AI. It saves compute, but cut too much and scores drop.
Model Compression
Network pruning is a common Model Compression method. It directly makes the network smaller.
Quantization
Quantization changes number precision. Pruning changes the network structure.
Parameter
Pruning often checks Parameter importance and removes weak weights.
Inference
After extra work is removed, Inference is often faster and lighter.