sparsity

Model pruning: removing weights without confusing file size with speed Lesson

Model pruning removes selected weights or network components, but real storage and speed gains depend on representation, hardware, and retained accuracy.