ResNet-18 + PACT + R2Loss

Reported on 3 benchmarks across 1 task · 1 paper

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

Methodology3 results

QuantizationonImageNet
Activation bits· 2023-03-14
4
best: 8 (FQ-ViT (ViT-L))
R2 Loss: Range Restriction Loss for Model Compression and Quantization arXiv:2303.08253
QuantizationonImageNet
Top-1 Accuracy (%)· 2023-03-14
68.45
best: 85.03 (FQ-ViT (ViT-L))
R2 Loss: Range Restriction Loss for Model Compression and Quantization arXiv:2303.08253
QuantizationonImageNet
Weight bits· 2023-03-14
2
best: 8 (FQ-ViT (ViT-L))
R2 Loss: Range Restriction Loss for Model Compression and Quantization arXiv:2303.08253