TensorFlow implementation of "ResNeSt: Split-Attention Networks"
ResNet-TF2
ResNeXt-TF2
WideResNet(WRN)-TF2
ResNet-with-LRWarmUp-TF2
ResNet-with-SGDR-TF2
Comparing SE-Net, SK-Net, and ResNeSt [1].
ResNeXt Remind
ResNeSt Diagram
Split-Attention in ResNest
| Indicator | Value |
|---|---|
| Accuracy | 0.99130 |
| Precision | 0.99123 |
| Recall | 0.99138 |
| F1-Score | 0.99129 |
Confusion Matrix
[[ 979 0 0 0 0 0 0 1 0 0]
[ 0 1118 0 3 2 0 2 5 5 0]
[ 0 1 1027 0 0 0 1 3 0 0]
[ 0 0 0 1007 0 2 0 0 1 0]
[ 0 0 0 0 977 0 2 0 0 3]
[ 0 0 0 3 0 885 1 1 1 1]
[ 7 1 0 0 0 1 949 0 0 0]
[ 0 2 3 0 0 0 0 1021 0 2]
[ 3 0 4 1 0 3 3 0 958 2]
[ 2 1 1 0 8 2 1 1 1 992]]
Class-0 | Precision: 0.98789, Recall: 0.99898, F1-Score: 0.99340
Class-1 | Precision: 0.99555, Recall: 0.98502, F1-Score: 0.99026
Class-2 | Precision: 0.99227, Recall: 0.99516, F1-Score: 0.99371
Class-3 | Precision: 0.99310, Recall: 0.99703, F1-Score: 0.99506
Class-4 | Precision: 0.98987, Recall: 0.99491, F1-Score: 0.99238
Class-5 | Precision: 0.99104, Recall: 0.99215, F1-Score: 0.99160
Class-6 | Precision: 0.98957, Recall: 0.99061, F1-Score: 0.99009
Class-7 | Precision: 0.98934, Recall: 0.99319, F1-Score: 0.99126
Class-8 | Precision: 0.99172, Recall: 0.98357, F1-Score: 0.98763
Class-9 | Precision: 0.99200, Recall: 0.98315, F1-Score: 0.98756
Total | Accuracy: 0.99130, Precision: 0.99123, Recall: 0.99138, F1-Score: 0.99129
- Python 3.7.6
- Tensorflow 2.1.0
- Numpy 1.18.1
- Matplotlib 3.1.3
[1] Hang Zhang et al. (2020). ResNeSt: Split-Attention Networks. arXiv preprint arXiv:2004.08955.
The authors of ResNeSt [1] have marked stars in this repository.
- Trained ResNest with MNIST dataset (ver. TF-Lite)