Resnet-50: 90% Sparse 100 epochs

Reported on 1 benchmark across 1 task · 2 papers

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Methodology2 results

Sparse LearningonImageNet
Top-1 Accuracy· 2021-06-19
74.5
best: 77.1 (Resnet-50: 80% Sparse)
Sparse Training via Boosting Pruning Plasticity with Neuroregeneration arXiv:2106.10404
Sparse LearningonImageNet
Top-1 Accuracy· 2021-02-04
73.82
best: 77.1 (Resnet-50: 80% Sparse)
Do We Actually Need Dense Over-Parameterization? In-Time Over-Parameterization in Sparse Training arXiv:2102.02887