File size: 7,121 Bytes
6c7aece c8b7f48 6c7aece 17b449d 6c7aece 17b449d 5fd43c0 17b449d 5fd43c0 17b449d 5fd43c0 17b449d 5fd43c0 40fae8e 5fd43c0 6c7aece | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 | ---
library_name: pytorch
license: other
tags:
- bu_auto
- real_time
- android
pipeline_tag: object-detection
---
# Yolo-R: Optimized for Qualcomm Devices
YoloR is a machine learning model that predicts bounding boxes and classes of objects in an image.
This is based on the implementation of Yolo-R found [here](https://github.com/WongKinYiu/yolor.git).
This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/qai_hub_models/models/yolor) 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
Due to licensing restrictions, we cannot distribute pre-exported model assets for this model.
Use the [Qualcomm® AI Hub Models](https://github.com/qualcomm/ai-hub-models/blob/main/qai_hub_models/models/yolor) 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
See our repository for [Yolo-R on GitHub](https://github.com/qualcomm/ai-hub-models/blob/main/qai_hub_models/models/yolor) for usage instructions.
## Model Details
**Model Type:** Model_use_case.object_detection
**Model Stats:**
- Model checkpoint: yolor_p6
- Input resolution: 640x640
- Number of parameters: 4.68M
- Model size (float): 17.9 MB
## Performance Summary
| Model | Runtime | Precision | Chipset | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit
|---|---|---|---|---|---|---
| Yolo-R | ONNX | float | Snapdragon® X2 Elite | 26.309 ms | 75 - 75 MB | NPU
| Yolo-R | ONNX | float | Snapdragon® X Elite | 57.577 ms | 74 - 74 MB | NPU
| Yolo-R | ONNX | float | Snapdragon® 8 Gen 3 Mobile | 41.986 ms | 2 - 346 MB | NPU
| Yolo-R | ONNX | float | Qualcomm® QCS8550 (Proxy) | 55.998 ms | 0 - 79 MB | NPU
| Yolo-R | ONNX | float | Qualcomm® QCS9075 | 52.751 ms | 5 - 12 MB | NPU
| Yolo-R | ONNX | float | Snapdragon® 8 Elite For Galaxy Mobile | 37.052 ms | 3 - 233 MB | NPU
| Yolo-R | ONNX | float | Snapdragon® 8 Elite Gen 5 Mobile | 26.443 ms | 6 - 318 MB | NPU
| Yolo-R | ONNX | w8a16 | Snapdragon® X2 Elite | 18.322 ms | 41 - 41 MB | NPU
| Yolo-R | ONNX | w8a16 | Snapdragon® X Elite | 30.49 ms | 40 - 40 MB | NPU
| Yolo-R | ONNX | w8a16 | Snapdragon® 8 Gen 3 Mobile | 20.897 ms | 0 - 490 MB | NPU
| Yolo-R | ONNX | w8a16 | Qualcomm® QCS6490 | 2282.662 ms | 129 - 136 MB | CPU
| Yolo-R | ONNX | w8a16 | Qualcomm® QCS8550 (Proxy) | 29.166 ms | 0 - 50 MB | NPU
| Yolo-R | ONNX | w8a16 | Qualcomm® QCS9075 | 29.553 ms | 1 - 6 MB | NPU
| Yolo-R | ONNX | w8a16 | Qualcomm® QCM6690 | 1176.534 ms | 48 - 62 MB | CPU
| Yolo-R | ONNX | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 17.119 ms | 1 - 364 MB | NPU
| Yolo-R | ONNX | w8a16 | Snapdragon® 7 Gen 4 Mobile | 1118.647 ms | 154 - 166 MB | CPU
| Yolo-R | ONNX | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 17.934 ms | 3 - 429 MB | NPU
| Yolo-R | QNN_DLC | float | Snapdragon® X2 Elite | 15.23 ms | 5 - 5 MB | NPU
| Yolo-R | QNN_DLC | float | Snapdragon® X Elite | 29.418 ms | 5 - 5 MB | NPU
| Yolo-R | QNN_DLC | float | Snapdragon® 8 Gen 3 Mobile | 22.761 ms | 0 - 317 MB | NPU
| Yolo-R | QNN_DLC | float | Qualcomm® QCS8275 (Proxy) | 100.329 ms | 0 - 252 MB | NPU
| Yolo-R | QNN_DLC | float | Qualcomm® QCS8550 (Proxy) | 29.567 ms | 5 - 7 MB | NPU
| Yolo-R | QNN_DLC | float | Qualcomm® SA8775P | 36.109 ms | 0 - 259 MB | NPU
| Yolo-R | QNN_DLC | float | Qualcomm® QCS9075 | 37.531 ms | 5 - 11 MB | NPU
| Yolo-R | QNN_DLC | float | Qualcomm® QCS8450 (Proxy) | 50.546 ms | 5 - 377 MB | NPU
| Yolo-R | QNN_DLC | float | Qualcomm® SA7255P | 100.329 ms | 0 - 252 MB | NPU
| Yolo-R | QNN_DLC | float | Qualcomm® SA8295P | 44.017 ms | 0 - 307 MB | NPU
| Yolo-R | QNN_DLC | float | Snapdragon® 8 Elite For Galaxy Mobile | 18.262 ms | 0 - 230 MB | NPU
| Yolo-R | QNN_DLC | float | Snapdragon® 8 Elite Gen 5 Mobile | 13.999 ms | 5 - 295 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Snapdragon® X2 Elite | 8.363 ms | 2 - 2 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Snapdragon® X Elite | 18.967 ms | 2 - 2 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Snapdragon® 8 Gen 3 Mobile | 12.323 ms | 2 - 370 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Qualcomm® QCS6490 | 77.324 ms | 3 - 7 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Qualcomm® QCS8275 (Proxy) | 37.394 ms | 0 - 288 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Qualcomm® QCS8550 (Proxy) | 18.116 ms | 2 - 4 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Qualcomm® SA8775P | 18.34 ms | 0 - 288 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Qualcomm® QCS9075 | 20.195 ms | 1 - 5 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Qualcomm® QCM6690 | 246.253 ms | 2 - 400 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Qualcomm® QCS8450 (Proxy) | 25.017 ms | 2 - 371 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Qualcomm® SA7255P | 37.394 ms | 0 - 288 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Qualcomm® SA8295P | 23.495 ms | 0 - 291 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Snapdragon® 8 Elite For Galaxy Mobile | 9.865 ms | 0 - 299 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Snapdragon® 7 Gen 4 Mobile | 26.414 ms | 2 - 312 MB | NPU
| Yolo-R | QNN_DLC | w8a16 | Snapdragon® 8 Elite Gen 5 Mobile | 7.295 ms | 2 - 303 MB | NPU
| Yolo-R | TFLITE | float | Snapdragon® 8 Gen 3 Mobile | 39.801 ms | 0 - 445 MB | NPU
| Yolo-R | TFLITE | float | Qualcomm® QCS8275 (Proxy) | 137.818 ms | 1 - 328 MB | NPU
| Yolo-R | TFLITE | float | Qualcomm® QCS8550 (Proxy) | 59.564 ms | 1 - 3 MB | NPU
| Yolo-R | TFLITE | float | Qualcomm® SA8775P | 66.658 ms | 1 - 310 MB | NPU
| Yolo-R | TFLITE | float | Qualcomm® QCS9075 | 50.464 ms | 1 - 86 MB | NPU
| Yolo-R | TFLITE | float | Qualcomm® QCS8450 (Proxy) | 88.543 ms | 1 - 471 MB | NPU
| Yolo-R | TFLITE | float | Qualcomm® SA7255P | 137.818 ms | 1 - 328 MB | NPU
| Yolo-R | TFLITE | float | Qualcomm® SA8295P | 76.899 ms | 1 - 355 MB | NPU
| Yolo-R | TFLITE | float | Snapdragon® 8 Elite For Galaxy Mobile | 31.258 ms | 1 - 281 MB | NPU
| Yolo-R | TFLITE | float | Snapdragon® 8 Elite Gen 5 Mobile | 26.051 ms | 0 - 341 MB | NPU
## License
* The license for the original implementation of Yolo-R can be found
[here](https://github.com/WongKinYiu/yolor/blob/main/LICENSE).
## References
* [You Only Learn One Representation: Unified Network for Multiple Tasks](https://arxiv.org/abs/2105.04206)
* [Source Model Implementation](https://github.com/WongKinYiu/yolor.git)
## 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).
|