EfficientNetV2: Smaller Models and Faster Training
Paper • 2104.00298 • Published • 1
EfficientNet-V2-s: Optimized for Qualcomm Devices
EfficientNetV2-s is a machine learning model that can classify images from the Imagenet dataset. It can also be used as a backbone in building more complex models for specific use cases.
This is based on the implementation of EfficientNet-V2-s found here. This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the Qualcomm® AI Hub Models library to export with custom configurations. More details on model performance across various devices, can be found here.
Qualcomm AI Hub Models uses Qualcomm AI Hub Workbench to compile, profile, and evaluate this model. Sign up to run these models on a hosted Qualcomm® device.
Getting Started
There are two ways to deploy this model on your device:
Option 1: Download Pre-Exported Models
Below are pre-exported model assets ready for deployment.
| Runtime | Precision | Chipset | SDK Versions | Download |
|---|---|---|---|---|
| ONNX | float | Universal | QAIRT 2.50, ONNX Runtime 1.30.0 | Download |
| ONNX | w8a16 | Universal | QAIRT 2.50, ONNX Runtime 1.30.0 | Download |
| QNN_DLC | float | Universal | QAIRT 2.50 | Download |
| QNN_DLC | w8a16 | Universal | QAIRT 2.50 | Download |
| TFLITE | float | Universal | QAIRT 2.50 | Download |
For more device-specific assets and performance metrics, visit EfficientNet-V2-s on Qualcomm® AI Hub.
Option 2: Export with Custom Configurations
Use the Qualcomm® AI Hub Models Python library to compile and export the model with your own:
- Custom weights (e.g., fine-tuned checkpoints)
- Custom input shapes
- Target device and runtime configurations
This option is ideal if you need to customize the model beyond the default configuration provided here.
See our repository for EfficientNet-V2-s on GitHub for usage instructions.
Model Details
Model Type: Model_use_case.image_classification
Model Stats:
- Input resolution: 384x384
- Model checkpoint: Imagenet
- Model size (float): 81.7 MB
- Model size (w8a16): 27.2 MB
- Number of parameters: 21.4M
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| EfficientNet-V2-s | ONNX | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.252 ms | 0 - 197 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.07 ms | 0 - 194 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Snapdragon® X2 Elite | 2.958 ms | 2 - 2 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Snapdragon® X Elite | 5.578 ms | 46 - 46 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 4.076 ms | 0 - 167 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 11.185 ms | 1 - 196 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Qualcomm® Dragonwing™ IQ-8275 | 8.09 ms | 2 - 7 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.304 ms | 0 - 209 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Qualcomm® QCS8450 | 11.185 ms | 1 - 196 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 7.621 ms | 1 - 6 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 5.578 ms | 46 - 46 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 3.07 ms | 0 - 194 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 1.996 ms | 0 - 165 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.54 ms | 0 - 163 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® X2 Elite | 2.326 ms | 2 - 2 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® X Elite | 5.683 ms | 24 - 24 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.656 ms | 0 - 205 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® 8 Gen 1 Mobile | 6.779 ms | 1 - 219 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 16.79 ms | 1 - 4 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 5.132 ms | 1 - 5 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.458 ms | 0 - 30 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® QCS8450 | 6.779 ms | 1 - 219 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 5.718 ms | 1 - 4 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 5.683 ms | 24 - 24 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 34.503 ms | 1 - 265 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 6.488 ms | 1 - 265 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 2.54 ms | 0 - 163 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 6.488 ms | 1 - 265 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.275 ms | 2 - 82 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.168 ms | 0 - 78 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® X2 Elite | 3.396 ms | 2 - 2 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® X Elite | 6.291 ms | 2 - 2 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 4.369 ms | 0 - 156 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 13.087 ms | 0 - 188 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-8275 | 8.19 ms | 2 - 6 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.715 ms | 2 - 143 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® SA8775P | 8.227 ms | 2 - 76 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® SA8650P | 8.227 ms | 2 - 76 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® SA8255P | 8.227 ms | 2 - 76 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® QCS8450 | 13.087 ms | 0 - 188 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 7.961 ms | 2 - 6 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 6.291 ms | 2 - 2 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 3.168 ms | 0 - 78 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® SA8295P | 12.93 ms | 2 - 108 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.196 ms | 1 - 153 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 2.842 ms | 0 - 150 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 2.882 ms | 1 - 1 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® X Elite | 6.794 ms | 1 - 1 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 4.2 ms | 0 - 181 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® 8 Gen 1 Mobile | 7.669 ms | 0 - 190 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 18.866 ms | 1 - 4 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-8275 | 5.795 ms | 1 - 5 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 6.191 ms | 1 - 3 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® SA8775P | 6.653 ms | 1 - 143 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® SA8650P | 6.653 ms | 1 - 143 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® SA8255P | 6.653 ms | 1 - 143 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® QCS8450 | 7.669 ms | 0 - 190 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 6.461 ms | 1 - 4 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 6.794 ms | 1 - 1 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 40.422 ms | 1 - 251 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 8.23 ms | 1 - 253 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 2.842 ms | 0 - 150 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® SA8295P | 8.023 ms | 1 - 141 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 8.23 ms | 1 - 253 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Snapdragon® 8 Elite Gen 5 For Galaxy Mobile | 2.334 ms | 0 - 122 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 3.152 ms | 0 - 123 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 4.384 ms | 0 - 200 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 13.19 ms | 0 - 230 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® Dragonwing™ IQ-8275 | 8.193 ms | 0 - 52 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.73 ms | 0 - 3 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® SA8775P | 8.19 ms | 0 - 119 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® SA8650P | 8.19 ms | 0 - 119 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® SA8255P | 8.19 ms | 0 - 119 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® QCS8450 | 13.19 ms | 0 - 230 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 7.938 ms | 0 - 51 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 3.152 ms | 0 - 123 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® SA7255P | 25.749 ms | 0 - 116 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® SA8295P | 12.863 ms | 0 - 149 MB | NPU |
License
- The license for the original implementation of EfficientNet-V2-s can be found here.
References
Community
- Join our AI Hub Slack community to collaborate, post questions and learn more about on-device AI.
- For questions or feedback please reach out to us.