notesum.ai
Published at December 9On How Iterative Magnitude Pruning Discovers Local Receptive Fields in Fully Connected Neural Networks
cs.LG
Released Date: December 9, 2024
Authors: William T. Redman1, Zhangyang Wang, Alessandro Ingrosso, Sebastian Goldt
Aff.: 1UC Santa Barbara

| Parameters | |
|---|---|
| Architecture | 3072:1024:1024:1000 |
| Activation function | ReLU |
| Batch size | 40 |
| Learning rate | 0.1 |
| Optimizer | SGD |
| Training iterations | 40,000 |
| Rewind iteration () | 1,000 |