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 2.45, ONNX Runtime 1.25.0 | Download |
| ONNX | w8a16 | Universal | QAIRT 2.45, ONNX Runtime 1.25.0 | Download |
| QNN_DLC | float | Universal | QAIRT 2.45 | Download |
| QNN_DLC | w8a16 | Universal | QAIRT 2.45 | Download |
| TFLITE | float | Universal | QAIRT 2.45 | 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:
- Model checkpoint: Imagenet
- Input resolution: 384x384
- Number of parameters: 21.4M
- Model size (float): 81.7 MB
- Model size (w8a16): 27.2 MB
Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|---|---|---|---|---|---|---|
| EfficientNet-V2-s | ONNX | float | Snapdragon® X2 Elite | 3.035 ms | 2 - 2 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Snapdragon® X Elite | 5.605 ms | 46 - 46 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 4.061 ms | 0 - 165 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Snapdragon® 8 Gen 1 Mobile | 11.635 ms | 2 - 197 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.347 ms | 0 - 49 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Qualcomm® QCS8450 | 11.635 ms | 2 - 197 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Qualcomm® Dragonwing™ IQ-9075 | 7.625 ms | 1 - 6 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.319 ms | 2 - 206 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Snapdragon® 8 Elite Mobile | 3.1 ms | 0 - 202 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Qualcomm® Dragonwing™ Q-8750 | 3.1 ms | 0 - 202 MB | NPU |
| EfficientNet-V2-s | ONNX | float | Qualcomm® Dragonwing™ IQ-X7181 | 5.605 ms | 46 - 46 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® X2 Elite | 2.365 ms | 2 - 2 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® X Elite | 5.739 ms | 24 - 24 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 3.622 ms | 0 - 208 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® 8 Gen 1 Mobile | 6.781 ms | 1 - 218 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 20.095 ms | 1 - 4 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.47 ms | 0 - 249 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® QCS8450 | 6.781 ms | 1 - 218 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 5.734 ms | 1 - 4 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 6.428 ms | 1 - 273 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 2.006 ms | 0 - 169 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 39.348 ms | 1 - 274 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Snapdragon® 8 Elite Mobile | 2.51 ms | 0 - 164 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 6.428 ms | 1 - 273 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 2.51 ms | 0 - 164 MB | NPU |
| EfficientNet-V2-s | ONNX | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 5.739 ms | 24 - 24 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® X2 Elite | 3.43 ms | 2 - 2 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® X Elite | 6.381 ms | 2 - 2 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 4.335 ms | 0 - 158 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® 8 Gen 1 Mobile | 13.484 ms | 0 - 188 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® QCS8275 | 25.737 ms | 2 - 76 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.782 ms | 2 - 3 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® SA8775P | 8.22 ms | 2 - 76 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® SA8650P | 8.22 ms | 2 - 76 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® SA8255P | 8.22 ms | 2 - 76 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® QCS8450 | 13.484 ms | 0 - 188 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-9075 | 7.911 ms | 4 - 7 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.361 ms | 2 - 85 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® SA7255P | 25.737 ms | 2 - 76 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Snapdragon® 8 Elite Mobile | 3.133 ms | 2 - 80 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® SA8295P | 12.985 ms | 2 - 108 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® Dragonwing™ Q-8750 | 3.133 ms | 2 - 80 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | float | Qualcomm® Dragonwing™ IQ-X7181 | 6.381 ms | 2 - 2 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 2.86 ms | 1 - 1 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® X Elite | 6.8 ms | 1 - 1 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 4.195 ms | 1 - 176 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® 8 Gen 1 Mobile | 8.155 ms | 1 - 185 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS6490 | 21.336 ms | 1 - 3 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® QCS8275 | 12.069 ms | 1 - 134 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 6.231 ms | 1 - 2 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® SA8775P | 6.7 ms | 1 - 137 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® SA8650P | 6.7 ms | 1 - 137 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® SA8255P | 6.7 ms | 1 - 137 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® QCS8450 | 8.155 ms | 1 - 185 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-9075 | 6.466 ms | 0 - 3 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 7.169 ms | 1 - 254 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 2.247 ms | 1 - 146 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-6690 | 41.217 ms | 1 - 255 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® SA7255P | 12.069 ms | 1 - 134 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Snapdragon® 8 Elite Mobile | 2.819 ms | 0 - 144 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® SA8295P | 8.38 ms | 1 - 135 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-7790 | 7.169 ms | 1 - 254 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ Q-8750 | 2.819 ms | 0 - 144 MB | NPU |
| EfficientNet-V2-s | QNN_DLC | w8a16 | Qualcomm® Dragonwing™ IQ-X7181 | 6.8 ms | 1 - 1 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 4.327 ms | 0 - 199 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Snapdragon® 8 Gen 1 Mobile | 13.528 ms | 0 - 228 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® QCS8275 | 25.814 ms | 0 - 116 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® Dragonwing™ QCS8550 (Proxy) | 5.74 ms | 0 - 2 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® SA8775P | 8.25 ms | 0 - 119 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® SA8650P | 8.25 ms | 0 - 119 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® SA8255P | 8.25 ms | 0 - 119 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® QCS8450 | 13.528 ms | 0 - 228 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® Dragonwing™ IQ-9075 | 7.914 ms | 0 - 51 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.339 ms | 0 - 121 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® SA7255P | 25.814 ms | 0 - 116 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Snapdragon® 8 Elite Mobile | 3.137 ms | 0 - 123 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® SA8295P | 13.029 ms | 0 - 148 MB | NPU |
| EfficientNet-V2-s | TFLITE | float | Qualcomm® Dragonwing™ Q-8750 | 3.137 ms | 0 - 123 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.