I designed a free parameter self-attention convolutional block module using pytorch.
My work is inspired from CBAM.
I tested my work on CIFAR-100 dataset and residual18 network.
| Model | Param. | Acc1. | Acc2. | Acc3. | Acc4. | Acc5. | Best Acc. | Avg Acc. |
|---|---|---|---|---|---|---|---|---|
| resnet18 | 11.22M | 76.36% | 75.94% | 76.38% | 76.03% | 76.37% | 76.38% | 76.22% |
| with CBAM | 11.39M | 76.20% | 76.55% | 76.23% | 76.26% | 76.16% | 76.55% | 76.28% |
| with ZCBAM(Max) | 11.22M | 75.64% | 75.97% | 76.20% | 75.99% | 75.87% | 76.20% | 75.93% |
| with ZCBAM(Avg) | 11.22M | 76.89% | 76.77% | 76.51% | 76.45% | 76.68% | 76.89% | 76.66% |
| with ZCBAM(Avg&Max) | 11.22M | 76.46% | 76.95% | 76.62% | 76.34% | 76.12% | 76.95% | 76.50% |
If you are interested in my work, please refer to
ZCBAM.
