Image Classification on MNIST (test) (Accuracy & Delta Metrics)
98.1Accuracyℓ1 magnitude pruning
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ℓ1 magnitude pruningSparsity=0%2026.06 | 98.1 | — | |
| CQPSparsity=0%2026.06 | 98.1 | — | |
| CQPSparsity=30%2026.06 | 98 | 0.2 | |
| CQPSparsity=50%2026.06 | 97.9 | 0.4 | |
| ℓ1 magnitude pruningSparsity=30%2026.06 | 97.8 | — | |
| ℓ1 magnitude pruningSparsity=50%2026.06 | 97.5 | — | |
| CQPSparsity=70%2026.06 | 97 | 0.9 | |
| ℓ1 magnitude pruningSparsity=70%2026.06 | 96.1 | — | |
| CQPSparsity=90%2026.06 | 95.6 | 2.2 | |
| ℓ1 magnitude pruningSparsity=90%2026.06 | 93.4 | — |