- Image Classification
Using automatic compression allows for easy model compression without a deep understanding of compression. Advanced compression offers higher flexibility in settings and may yield better compression results.
- 'Compressed' Type: Automatic Compression
- 'Compressed(Adv.)' Type: Advanced Compression
All of the original and compressed models can be downloaded easily on the Model Compressor Model Zoo.
PyTorch
| Model | Best Practice | Type | Dataset | Accuracy (%) | FLOPs (M) | Params (M) | Latency (ms) | Model Size (MB) |
|---|---|---|---|---|---|---|---|---|
| VGG16 | Original | CIFAR100 | 74.00 | 629.76 | 15.30 | 71.65 | 59.65 | |
| VGG16 | Google Colab | Compressed-1 | CIFAR100 | 72.22 (-1.78) | 431.84 (1.46x) | 5.16 (2.96x) | 24.52 (2.91x) | 20.22 (2.95x) |
| VGG16 | Google Colab | Compressed-2 | CIFAR100 | 68.01 (-5.99) | 213.06 (2.96x) | 1.25 (12.26x) | 11.34 (6.32x) | 4.93 (12.10x) |
| MobileNetV2 | Original | CIFAR100 | 74.29 | 189.30 | 2.35 | 46.26 | 8.98 | |
| MobileNetV2 | Google Colab | Compressed | CIFAR100 | 73.68 (-0.61) | 119.09 (1.59x) | 0.82 (2.88x) | 24.50 (1.89x) | 3.38 (2.66x) |
| RepVGG | Original | CIFAR100 | 76.44 | 1715.70 | 12.94 | 248.10 | 50.33 | |
| RepVGG | Google Colab | Compressed-1 | CIFAR100 | 74.92 (-1.52) | 1644.88 (1.04x) | 10.64 (1.22x) | 113.35 (2.19x) | 41.81 (1.20x) |
| RepVGG | Google Colab | Compressed-2 | CIFAR100 | 69.84 (-4.60) | 721.77 (2.38x) | 2.95 (4.39x) | 51.69 (4.80x) | 11.71 (4.30x) |
| ViT | Original | CIFAR100 | 94.42 | 33725.76 | 85.80 | 1396.53 | 327.43 | |
| ViT | Google Colab | Compressed | CIFAR100 | 93.30 (-1.12) | 14804.95 (2.28x) | 37.78 (2.27x) | 737.11 (1.89x) | 144.32 (2.27x) |
- The model’s latency is measured on Raspberry Pi 4B (1.5GHz ARM Cortex).
- Options: FP32, ONNX runtime
TensorFlow-Keras
| Model | Best Practice | Type | Dataset | Accuracy (%) | FLOPs (M) | Params (M) | Latency (ms) | Model Size (MB) |
|---|---|---|---|---|---|---|---|---|
| VGG19 | Original | CIFAR-100 | 72.28 | 796.79 | 20.09 | 189.31 | 78.69 | |
| VGG19 | Google Colab | Compressed | CIFAR-100 | 71.13 (-1.15) | 132.20 (6.03x) | 1.17 (17.13x) | 12.85 (14.73x) | 4.98 (15.80x) |
| VGG19 | Compressed (Adv.) | CIFAR-100 | 71.14 (-1.14) | 100.09 (7.96x) | 0.66 (30.38x) | 4.5 (42.06x) | 5.68 (13.85x) | |
| ResNet50 | Original | CIFAR-100 | 78.03 | 2596.06 | 23.71 | 450.14 | 93.31 | |
| ResNet50 | Google Colab | Compressed | CIFAR-100 | 76.92 (-1.11) | 613.43 (4.23x) | 2.64 (8.99x) | 130.39 (3.45x) | 9.83 (9.49x) |
| ResNet50 | Compressed (Adv.) | CIFAR-100 | 76.63 (-1.4) | 224.70 (11.55x) | 2.17 (10.91x) | 48.37 (9.31x) | 18.35 (5.09x) | |
| MobileNet V1 | Original | CIFAR-100 | 66.68 | 92.90 | 3.31 | 35.61 | 13.28 | |
| MobileNet V1 | Google Colab | Compressed | CIFAR-100 | 66.32 (-0.36) | 26.09 (3.56x) | 0.53 (6.24x) | 3.66 (9.73x) | 2.38 (5.58x) |
| MobileNet V1 | Compressed (Adv.) | CIFAR-100 | 66.11 (-0.57) | 17.90 (5.19x) | 0.35 (9.35x) | 2.08 (17.12x) | 3.3 (4.02x) |
- The model’s latency is measured on Raspberry Pi 4B (1.5GHz ARM Cortex).
- Options: FP32, TensorFlow Lite
Updated 25 days ago