File size: 6,440 Bytes
dc1bcb4 9ce8f8b dc1bcb4 5ca5b6d dc1bcb4 13426a5 dc1bcb4 88eba9c 13426a5 0cef771 dc1bcb4 5c1b282 dc1bcb4 be78927 5c1b282 dc1bcb4 c6bcf71 dc1bcb4 be78927 8ac63c2 be78927 dc1bcb4 be78927 dc1bcb4 be78927 dc1bcb4 be78927 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 |
---
library_name: pytorch
license: other
tags:
- real_time
- quantized
- android
pipeline_tag: object-detection
---
# Yolo-NAS-Quantized: Optimized for Mobile Deployment
## Quantized real-time object detection optimized for mobile and edge
YoloNAS is a machine learning model that predicts bounding boxes and classes of objects in an image. This model is post-training quantized to int8 using samples from the COCO dataset.
This model is an implementation of Yolo-NAS-Quantized found [here](https://github.com/Deci-AI/super-gradients).
More details on model performance across various devices, can be found [here](https://aihub.qualcomm.com/models/yolonas_quantized).
### Model Details
- **Model Type:** Object detection
- **Model Stats:**
- Model checkpoint: YoloNAS Small
- Input resolution: 640x640
- Number of parameters: 12.2M
- Model size: 12.1 MB
- Precision: w8a8 (8-bit weights, 8-bit activations)
| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model
|---|---|---|---|---|---|---|---|---|
| Yolo-NAS-Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 3.587 ms | 0 - 23 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 3.487 ms | 1 - 3 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 16.196 ms | 1 - 62 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 2.389 ms | 0 - 48 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 2.409 ms | 1 - 20 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 12.8 ms | 4 - 190 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 1.831 ms | 0 - 37 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 2.212 ms | 1 - 36 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 10.914 ms | 7 - 182 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | SA7255P ADP | SA7255P | TFLITE | 32.217 ms | 0 - 33 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | SA7255P ADP | SA7255P | QNN | 31.979 ms | 1 - 11 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 3.555 ms | 0 - 23 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | SA8255 (Proxy) | SA8255P Proxy | QNN | 3.485 ms | 1 - 3 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | SA8295P ADP | SA8295P | TFLITE | 5.351 ms | 0 - 35 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | SA8295P ADP | SA8295P | QNN | 5.264 ms | 1 - 19 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 3.538 ms | 0 - 19 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | SA8650 (Proxy) | SA8650P Proxy | QNN | 3.488 ms | 1 - 4 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | SA8775P ADP | SA8775P | TFLITE | 4.992 ms | 0 - 33 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | SA8775P ADP | SA8775P | QNN | 4.83 ms | 1 - 11 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | TFLITE | 11.03 ms | 0 - 46 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | RB3 Gen 2 (Proxy) | QCS6490 Proxy | QNN | 15.09 ms | 1 - 15 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | QCS8275 (Proxy) | QCS8275 Proxy | TFLITE | 32.217 ms | 0 - 33 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | QCS8275 (Proxy) | QCS8275 Proxy | QNN | 31.979 ms | 1 - 11 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 3.568 ms | 0 - 20 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 3.488 ms | 1 - 3 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | QCS9075 (Proxy) | QCS9075 Proxy | TFLITE | 4.992 ms | 0 - 33 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | QCS9075 (Proxy) | QCS9075 Proxy | QNN | 4.83 ms | 1 - 11 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 4.267 ms | 0 - 50 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 4.705 ms | 1 - 43 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 3.818 ms | 1 - 1 MB | INT8 | NPU | -- |
| Yolo-NAS-Quantized | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 16.824 ms | 14 - 14 MB | INT8 | NPU | -- |
## License
* The license for the original implementation of Yolo-NAS-Quantized can be found
[here](https://github.com/Deci-AI/super-gradients/blob/master/YOLONAS.md#license).
* The license for the compiled assets for on-device deployment can be found [here](https://github.com/Deci-AI/super-gradients/blob/master/LICENSE.YOLONAS.md)
## References
* [YOLO-NAS by Deci Achieves SOTA Performance on Object Detection Using Neural Architecture Search](https://deci.ai/blog/yolo-nas-object-detection-foundation-model/)
* [Source Model Implementation](https://github.com/Deci-AI/super-gradients)
## Community
* Join [our AI Hub Slack community](https://qualcomm-ai-hub.slack.com/join/shared_invite/zt-2d5zsmas3-Sj0Q9TzslueCjS31eXG2UA#/shared-invite/email) to collaborate, post questions and learn more about on-device AI.
* For questions or feedback please [reach out to us](mailto:[email protected]).
## Usage and Limitations
Model may not be used for or in connection with any of the following applications:
- Accessing essential private and public services and benefits;
- Administration of justice and democratic processes;
- Assessing or recognizing the emotional state of a person;
- Biometric and biometrics-based systems, including categorization of persons based on sensitive characteristics;
- Education and vocational training;
- Employment and workers management;
- Exploitation of the vulnerabilities of persons resulting in harmful behavior;
- General purpose social scoring;
- Law enforcement;
- Management and operation of critical infrastructure;
- Migration, asylum and border control management;
- Predictive policing;
- Real-time remote biometric identification in public spaces;
- Recommender systems of social media platforms;
- Scraping of facial images (from the internet or otherwise); and/or
- Subliminal manipulation
|