Image Classification on ImageNet 1K (test) with FLOPs Reporting
85.8Top-1 AccuracyFasterViT-6
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| FasterViT-6Image Size (Px)=224, #Param (M)=1360.0, FLOPs (G)=142.0, Throughput (Img/Sec)=3522023.06 | 85.8 | — | — | — | |
| FasterViT-5Image Size (Px)=224, #Param (M)=957.5, FLOPs (G)=113.0, Throughput (Img/Sec)=4492023.06 | 85.6 | — | — | — | |
| FasterViT-4Image Size (Px)=224, #Param (M)=424.6, FLOPs (G)=36.6, Throughput (Img/Sec)=8492023.06 | 85.4 | — | — | — | |
| MaxViT-LImage Size (Px)=224, #Param (M)=212.0, FLOPs (G)=43.9, Throughput (Img/Sec)=3762023.06 | 85.1 | — | — | — | |
| MaxViT-BImage Size (Px)=224, #Param (M)=120.0, FLOPs (G)=23.4, Throughput (Img/Sec)=5072023.06 | 84.9 | — | — | — | |
| FasterViT-3Image Size (Px)=224, #Param (M)=159.5, FLOPs (G)=18.2, Throughput (Img/Sec)=17802023.06 | 84.9 | — | — | — | |
| DeiT3-LImage Size (Px)=224, #Param (M)=304.4, FLOPs (G)=59.7, Throughput (Img/Sec)=5352023.06 | 84.8 | — | — | — | |
| SwinV2-BImage Size (Px)=256, #Param (M)=87.9, FLOPs (G)=15.1, Throughput (Img/Sec)=5352023.06 | 84.6 | — | — | — | |
| FasterViT-2Image Size (Px)=224, #Param (M)=75.9, FLOPs (G)=8.7, Throughput (Img/Sec)=31612023.06 | 84.2 | — | — | — | |
| EfficientNetV2-SImage Size (Px)=384, #Param (M)=21.5, FLOPs (G)=8.0, Throughput (Img/Sec)=17352023.06 | 83.9 | — | — | — | |
| ConvNeXt-BImage Size (Px)=224, #Param (M)=88.6, FLOPs (G)=15.4, Throughput (Img/Sec)=14852023.06 | 83.8 | — | — | — | |
| SwinV2-SImage Size (Px)=256, #Param (M)=49.7, FLOPs (G)=8.5, Throughput (Img/Sec)=10432023.06 | 83.8 | — | — | — | |
| EfficientFormer-L7Image Size (Px)=224, #Param (M)=82.2, FLOPs (G)=10.2, Throughput (Img/Sec)=13592023.06 | 83.4 | — | — | — | |
| Swin-SImage Size (Px)=224, #Param (M)=49.6, FLOPs (G)=8.5, Throughput (Img/Sec)=17202023.06 | 83.2 | — | — | — | |
| FasterViT-1Image Size (Px)=224, #Param (M)=53.4, FLOPs (G)=5.3, Throughput (Img/Sec)=41882023.06 | 83.2 | — | — | — | |
| ConvNeXt-SImage Size (Px)=224, #Param (M)=50.2, FLOPs (G)=8.7, Throughput (Img/Sec)=20082023.06 | 83.1 | — | — | — | |
| Twins-BImage Size (Px)=224, #Param (M)=56.1, FLOPs (G)=8.3, Throughput (Img/Sec)=19262023.06 | 83.1 | — | — | — | |
| RegNetY-040Image Size (Px)=288, #Param (M)=20.6, FLOPs (G)=6.6, Throughput (Img/Sec)=32272023.06 | 83 | — | — | — | |
| PoolFormer-M58Image Size (Px)=224, #Param (M)=73.5, FLOPs (G)=11.6, Throughput (Img/Sec)=8842023.06 | 82.4 | — | — | — | |
| CoaT-Lite-SImage Size (Px)=224, #Param (M)=19.8, FLOPs (G)=4.1, Throughput (Img/Sec)=22692023.06 | 82.3 | — | — | — | |
| CrossViT-BImage Size (Px)=240, #Param (M)=105.0, FLOPs (G)=20.1, Throughput (Img/Sec)=13212023.06 | 82.2 | — | — | — | |
| Visformer-SImage Size (Px)=224, #Param (M)=40.2, FLOPs (G)=4.8, Throughput (Img/Sec)=36762023.06 | 82.1 | — | — | — | |
| FasterViT-0Image Size (Px)=224, #Param (M)=31.4, FLOPs (G)=3.3, Throughput (Img/Sec)=58022023.06 | 82.1 | — | — | — | |
| ConvNeXt-TImage Size (Px)=224, #Param (M)=28.6, FLOPs (G)=4.5, Throughput (Img/Sec)=31962023.06 | 82 | — | — | — | |
| ResNetV2-101Image Size (Px)=224, #Param (M)=44.5, FLOPs (G)=7.8, Throughput (Img/Sec)=40192023.06 | 82 | — | — | — | |
| SwinV2-TImage Size (Px)=256, #Param (M)=28.3, FLOPs (G)=4.4, Throughput (Img/Sec)=16742023.06 | 81.8 | — | — | — | |
| Swin-TImage Size (Px)=224, #Param (M)=28.3, FLOPs (G)=4.4, Throughput (Img/Sec)=27582023.06 | 81.3 | — | — | — | |
| EdgeViT-SImage Size (Px)=224, #Param (M)=13.1, FLOPs (G)=1.9, Throughput (Img/Sec)=42542023.06 | 81 | — | — | — | |
| ResNet-50 (Dense)Sparsity=0%, Training Epochs=100, Backbone=ResNet-502023.05 | 76.8 | 1 | 1 | 1 | |
| ChaseSparsity=80%, Training Epochs=150, Channel Sparsity (Sc)=0, Backbone=ResNet-502023.05 | 76.67 | 0.55 | 0.34 | 1 | |
| ChaseSparsity=80%, Training Epochs=150, Channel Sparsity (Sc)=0.3, Backbone=ResNet-502023.05 | 76.23 | 0.57 | 0.36 | 0.75 | |
| ChaseSparsity=80%, Training Epochs=150, Channel Sparsity (Sc)=0.4, Backbone=ResNet-502023.05 | 76 | 0.59 | 0.37 | 0.68 | |
| ChaseSparsity=80%, Training Epochs=100, Channel Sparsity (Sc)=0, Backbone=ResNet-502023.05 | 75.87 | 0.37 | 0.34 | 1 | |
| RigL-ITOPSparsity=80%, Training Epochs=100, Backbone=ResNet-502023.05 | 75.84 | 0.42 | 0.42 | 1 | |
| ChaseSparsity=90%, Training Epochs=150, Channel Sparsity (Sc)=0, Backbone=ResNet-502023.05 | 75.77 | 0.36 | 0.21 | 1 | |
| MESTSparsity=80%, Training Epochs=150, Backbone=ResNet-502023.05 | 75.73 | 0.4 | 0.21 | 1 | |
| ChaseSparsity=80%, Training Epochs=100, Channel Sparsity (Sc)=0.3, Backbone=ResNet-502023.05 | 75.62 | 0.39 | 0.36 | 0.75 | |
| MESTSparsity=80%, Training Epochs=100, Backbone=ResNet-502023.05 | 75.39 | 0.23 | 0.21 | 1 | |
| ChaseSparsity=80%, Training Epochs=100, Channel Sparsity (Sc)=0.4, Backbone=ResNet-502023.05 | 75.27 | 0.39 | 0.37 | 0.68 | |
| SNFSSparsity=80%, Training Epochs=100, Backbone=ResNet-502023.05 | 75.2 | 0.61 | 0.42 | 1 | |
| ChaseSparsity=90%, Training Epochs=150, Channel Sparsity (Sc)=0.3, Backbone=ResNet-502023.05 | 75.2 | 0.37 | 0.22 | 0.74 | |
| RigLSparsity=80%, Training Epochs=100, Backbone=ResNet-502023.05 | 75.1 | 0.42 | 0.42 | 1 | |
| MESTSparsity=90%, Training Epochs=150, Backbone=ResNet-502023.05 | 75 | 0.2 | 0.11 | 1 | |
| ChaseSparsity=90%, Training Epochs=150, Channel Sparsity (Sc)=0.4, Backbone=ResNet-502023.05 | 74.87 | 0.38 | 0.23 | 0.67 | |
| ChaseSparsity=90%, Training Epochs=100, Channel Sparsity (Sc)=0, Backbone=ResNet-502023.05 | 74.7 | 0.24 | 0.21 | 1 | |
| ChaseSparsity=90%, Training Epochs=100, Channel Sparsity (Sc)=0.3, Backbone=ResNet-502023.05 | 74.35 | 0.25 | 0.22 | 0.74 | |
| ChaseSparsity=90%, Training Epochs=100, Channel Sparsity (Sc)=0.4, Backbone=ResNet-502023.05 | 74.03 | 0.26 | 0.23 | 0.67 | |
| RigL-ITOPSparsity=90%, Training Epochs=100, Backbone=ResNet-502023.05 | 73.82 | 0.25 | 0.24 | 1 | |
| DSRSparsity=80%, Training Epochs=100, Backbone=ResNet-502023.05 | 73.3 | 0.4 | 0.4 | 1 | |
| RigLSparsity=90%, Training Epochs=100, Backbone=ResNet-502023.05 | 73 | 0.25 | 0.24 | 1 | |
| SETSparsity=80%, Training Epochs=100, Backbone=ResNet-502023.05 | 72.9 | 0.23 | 0.23 | 1 | |
| SNFSSparsity=90%, Training Epochs=100, Backbone=ResNet-502023.05 | 72.9 | 0.5 | 0.24 | 1 | |
| MESTSparsity=90%, Training Epochs=100, Backbone=ResNet-502023.05 | 72.58 | 0.12 | 0.11 | 1 | |
| DSRSparsity=90%, Training Epochs=100, Backbone=ResNet-502023.05 | 71.6 | 0.3 | 0.3 | 1 | |
| SETSparsity=90%, Training Epochs=100, Backbone=ResNet-502023.05 | 69.6 | 0.1 | 0.1 | 1 |