Stroke Type Classification on Stroke type classification ischemia vs. hemorrhage (test)
98.8AccuracyStrokeNeXt-base
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| StrokeNeXt-baseLatency (s)=0.005, Throughput (img/s)=219, Peak GPU (GB)=2.641, FLOPs (G)=30.853, Params (M)=178.5, Train. time (h)=0.1012026.02 | 98.8 | 98.8 | 98.8 | 98.8 | 1 | 1 | 98.6 | 0.973 | 0.011 | 0.054 | |
| StrokeNeXt-tinyLatency (s)=0.002, Throughput (img/s)=571, Peak GPU (GB)=1.417, FLOPs (G)=8.977, Params (M)=57.6, Train. time (h)=0.0412026.02 | 98.6 | 98.5 | 98.6 | 98.6 | 1 | 1 | 98.6 | 0.973 | 0.011 | 0.058 | |
| VGG16Latency (s)=0.001, Throughput (img/s)=1434, Peak GPU (GB)=4.020, FLOPs (G)=15.519, Params (M)=134.3, Train. time (h)=0.0202026.02 | 87.9 | 87.9 | 87.9 | 87.9 | 0.951 | 0.956 | 87.9 | 0.758 | 0.104 | 0.124 | |
| ResNet152Latency (s)=0.001, Throughput (img/s)=1217, Peak GPU (GB)=1.298, FLOPs (G)=11.601, Params (M)=58.2, Train. time (h)=0.0252026.02 | 86.5 | 87.3 | 86.5 | 86.5 | 0.949 | 0.936 | 86.4 | 0.738 | 0.101 | 0.13 | |
| ResNet50Latency (s)=0.001, Throughput (img/s)=1900, Peak GPU (GB)=1.034, FLOPs (G)=4.130, Params (M)=23.5, Train. time (h)=0.0242026.02 | 85.2 | 85.4 | 85.2 | 85.2 | 0.918 | 0.924 | 85.1 | 0.706 | 0.133 | 0.136 | |
| ConvNeXt-baseLatency (s)=0.002, Throughput (img/s)=441, Peak GPU (GB)=1.947, FLOPs (G)=15.425, Params (M)=87.6, Train. time (h)=0.0262026.02 | 84.8 | 84.8 | 84.8 | 84.8 | 0.926 | 0.93 | 84.8 | 0.695 | 0.13 | 0.142 | |
| Swin TransformerLatency (s)=0.003, Throughput (img/s)=335, Peak GPU (GB)=2.196, FLOPs (G)=10.550, Params (M)=87.1, Train. time (h)=0.0292026.02 | 82.5 | 82.5 | 82.5 | 82.5 | 0.905 | 0.911 | 82.5 | 0.65 | 0.132 | 0.084 | |
| MobileNetv2Latency (s)=0.001, Throughput (img/s)=2341, Peak GPU (GB)=0.868, FLOPs (G)=0.319, Params (M)=2.2, Train. time (h)=0.0202026.02 | 81.2 | 81.3 | 81.1 | 81.1 | 0.919 | 0.919 | 81.1 | 0.624 | 0.122 | 0.083 |