|
|
--- |
|
|
library_name: pytorch |
|
|
license: other |
|
|
tags: |
|
|
- backbone |
|
|
- bu_auto |
|
|
- real_time |
|
|
- android |
|
|
pipeline_tag: image-classification |
|
|
|
|
|
--- |
|
|
|
|
|
 |
|
|
|
|
|
# EfficientViT-b2-cls: Optimized for Qualcomm Devices |
|
|
|
|
|
EfficientViT 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 EfficientViT-b2-cls found [here](https://github.com/CVHub520/efficientvit). |
|
|
This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/efficientvit_b2_cls) library to export with custom configurations. More details on model performance across various devices, can be found [here](#performance-summary). |
|
|
|
|
|
Qualcomm AI Hub Models uses [Qualcomm AI Hub Workbench](https://workbench.aihub.qualcomm.com) to compile, profile, and evaluate this model. [Sign up](https://myaccount.qualcomm.com/signup) 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 2.37, ONNX Runtime 1.23.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/efficientvit_b2_cls/releases/v0.46.0/efficientvit_b2_cls-onnx-float.zip) |
|
|
| ONNX | w8a16 | Universal | QAIRT 2.37, ONNX Runtime 1.23.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/efficientvit_b2_cls/releases/v0.46.0/efficientvit_b2_cls-onnx-w8a16.zip) |
|
|
| QNN_DLC | float | Universal | QAIRT 2.42 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/efficientvit_b2_cls/releases/v0.46.0/efficientvit_b2_cls-qnn_dlc-float.zip) |
|
|
| TFLITE | float | Universal | QAIRT 2.42, TFLite 2.17.0 | [Download](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/efficientvit_b2_cls/releases/v0.46.0/efficientvit_b2_cls-tflite-float.zip) |
|
|
|
|
|
For more device-specific assets and performance metrics, visit **[EfficientViT-b2-cls on Qualcomm® AI Hub](https://aihub.qualcomm.com/models/efficientvit_b2_cls)**. |
|
|
|
|
|
|
|
|
### Option 2: Export with Custom Configurations |
|
|
|
|
|
Use the [Qualcomm® AI Hub Models](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/efficientvit_b2_cls) 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 [EfficientViT-b2-cls on GitHub](https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/models/efficientvit_b2_cls) for usage instructions. |
|
|
|
|
|
## Model Details |
|
|
|
|
|
**Model Type:** Model_use_case.image_classification |
|
|
|
|
|
**Model Stats:** |
|
|
- Model checkpoint: Imagenet |
|
|
- Input resolution: 224x224 |
|
|
- Number of parameters: 24.3M |
|
|
- Model size (float): 92.9 MB |
|
|
|
|
|
## Performance Summary |
|
|
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit |
|
|
|---|---|---|---|---|---|--- |
|
|
| EfficientViT-b2-cls | ONNX | float | Snapdragon® X Elite | 5.878 ms | 49 - 49 MB | NPU |
|
|
| EfficientViT-b2-cls | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 3.748 ms | 0 - 222 MB | NPU |
|
|
| EfficientViT-b2-cls | ONNX | float | Qualcomm® QCS8550 (Proxy) | 5.425 ms | 0 - 58 MB | NPU |
|
|
| EfficientViT-b2-cls | ONNX | float | Qualcomm® QCS9075 | 5.902 ms | 1 - 4 MB | NPU |
|
|
| EfficientViT-b2-cls | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.858 ms | 0 - 151 MB | NPU |
|
|
| EfficientViT-b2-cls | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.466 ms | 0 - 151 MB | NPU |
|
|
| EfficientViT-b2-cls | QNN_DLC | float | Snapdragon® X Elite | 6.192 ms | 1 - 1 MB | NPU |
|
|
| EfficientViT-b2-cls | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 3.789 ms | 0 - 164 MB | NPU |
|
|
| EfficientViT-b2-cls | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 12.91 ms | 1 - 90 MB | NPU |
|
|
| EfficientViT-b2-cls | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 5.425 ms | 1 - 166 MB | NPU |
|
|
| EfficientViT-b2-cls | QNN_DLC | float | Qualcomm® QCS9075 | 6.198 ms | 1 - 3 MB | NPU |
|
|
| EfficientViT-b2-cls | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 7.172 ms | 0 - 165 MB | NPU |
|
|
| EfficientViT-b2-cls | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.797 ms | 1 - 92 MB | NPU |
|
|
| EfficientViT-b2-cls | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.335 ms | 1 - 95 MB | NPU |
|
|
| EfficientViT-b2-cls | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 3.795 ms | 0 - 224 MB | NPU |
|
|
| EfficientViT-b2-cls | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 12.956 ms | 0 - 148 MB | NPU |
|
|
| EfficientViT-b2-cls | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 5.468 ms | 0 - 3 MB | NPU |
|
|
| EfficientViT-b2-cls | TFLITE | float | Qualcomm® QCS9075 | 6.169 ms | 0 - 52 MB | NPU |
|
|
| EfficientViT-b2-cls | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 7.164 ms | 0 - 226 MB | NPU |
|
|
| EfficientViT-b2-cls | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 2.803 ms | 0 - 140 MB | NPU |
|
|
| EfficientViT-b2-cls | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 2.331 ms | 0 - 142 MB | NPU |
|
|
|
|
|
## License |
|
|
* The license for the original implementation of EfficientViT-b2-cls can be found |
|
|
[here](https://github.com/CVHub520/efficientvit/blob/main/LICENSE). |
|
|
|
|
|
## References |
|
|
* [EfficientViT: Multi-Scale Linear Attention for High-Resolution Dense Prediction](https://arxiv.org/abs/2205.14756) |
|
|
* [Source Model Implementation](https://github.com/CVHub520/efficientvit) |
|
|
|
|
|
## Community |
|
|
* Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI. |
|
|
* For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com). |
|
|
|
