Yolo-R / README.md

v0.59.0

ab8984c verified 2 days ago

6.3 kB

	---
	library_name: pytorch
	license: other
	tags:
	- bu_auto
	- real_time
	- android
	pipeline_tag: object-detection

	---

	![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/yolor/web-assets/model_demo.png)

	# 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).
	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/v0.59.0/src/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/v0.59.0/src/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/v0.59.0/src/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 \| 26.79 ms \| 5 - 5 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Snapdragon® X Elite \| 40.751 ms \| 75 - 75 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Snapdragon® 8 Gen 3 Mobile \| 31.447 ms \| 0 - 331 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Snapdragon® 8 Gen 1 Mobile \| 92.147 ms \| 6 - 524 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Qualcomm® Dragonwing™ QCS8550 (Proxy) \| 40.026 ms \| 5 - 8 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Qualcomm® QCS8450 \| 92.147 ms \| 6 - 524 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Qualcomm® Dragonwing™ IQ-9075 \| 47.694 ms \| 5 - 12 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Qualcomm® Dragonwing™ IQ-X7181 \| 40.751 ms \| 75 - 75 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Qualcomm® Dragonwing™ Q-8750 \| 28.044 ms \| 3 - 233 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Snapdragon® 8 Elite Mobile \| 28.044 ms \| 3 - 233 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Snapdragon® 8 Elite Gen 5 Mobile \| 27.585 ms \| 3 - 295 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® X2 Elite \| 19.175 ms \| 2 - 2 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® X Elite \| 30.831 ms \| 41 - 41 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® 8 Gen 3 Mobile \| 20.922 ms \| 1 - 464 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® 8 Gen 1 Mobile \| 42.424 ms \| 3 - 461 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Qualcomm® Dragonwing™ QCS8550 (Proxy) \| 29.41 ms \| 2 - 6 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Qualcomm® QCS8450 \| 42.424 ms \| 3 - 461 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Qualcomm® Dragonwing™ IQ-9075 \| 30.957 ms \| 1 - 6 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Qualcomm® Dragonwing™ IQ-X7181 \| 30.831 ms \| 41 - 41 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Qualcomm® Dragonwing™ Q-8750 \| 17.867 ms \| 2 - 365 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® 8 Elite Mobile \| 17.867 ms \| 2 - 365 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® 8 Elite Gen 5 Mobile \| 18.762 ms \| 1 - 425 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® X2 Elite \| 8.641 ms \| 2 - 2 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® X Elite \| 19.509 ms \| 2 - 2 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® 8 Gen 3 Mobile \| 12.447 ms \| 2 - 356 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® 8 Gen 1 Mobile \| 28.046 ms \| 2 - 358 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ QCS6490 \| 91.189 ms \| 2 - 7 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ QCS8275 \| 37.802 ms \| 2 - 291 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ QCS8550 (Proxy) \| 18.513 ms \| 2 - 4 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® SA8775P \| 18.744 ms \| 1 - 291 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® SA8650P \| 18.744 ms \| 1 - 291 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® SA8255P \| 18.744 ms \| 1 - 291 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® QCS8450 \| 28.046 ms \| 2 - 358 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ IQ-9075 \| 18.861 ms \| 0 - 4 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ IQ-X7181 \| 19.509 ms \| 2 - 2 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ Q-6690 \| 219.672 ms \| 2 - 399 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ Q-7790 \| 25.965 ms \| 2 - 309 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ Q-8750 \| 9.391 ms \| 2 - 304 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® SA7255P \| 37.802 ms \| 2 - 291 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® SA8295P \| 24.277 ms \| 0 - 292 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® 8 Elite Mobile \| 9.391 ms \| 2 - 304 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® 8 Elite Gen 5 Mobile \| 7.859 ms \| 2 - 310 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® 7 Gen 4 Mobile \| 25.965 ms \| 2 - 309 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)

	## 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).

	---
	library_name: pytorch
	license: other
	tags:
	- bu_auto
	- real_time
	- android
	pipeline_tag: object-detection

	---

	![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/yolor/web-assets/model_demo.png)

	# 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).
	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/v0.59.0/src/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/v0.59.0/src/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/v0.59.0/src/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 \| 26.79 ms \| 5 - 5 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Snapdragon® X Elite \| 40.751 ms \| 75 - 75 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Snapdragon® 8 Gen 3 Mobile \| 31.447 ms \| 0 - 331 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Snapdragon® 8 Gen 1 Mobile \| 92.147 ms \| 6 - 524 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Qualcomm® Dragonwing™ QCS8550 (Proxy) \| 40.026 ms \| 5 - 8 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Qualcomm® QCS8450 \| 92.147 ms \| 6 - 524 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Qualcomm® Dragonwing™ IQ-9075 \| 47.694 ms \| 5 - 12 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Qualcomm® Dragonwing™ IQ-X7181 \| 40.751 ms \| 75 - 75 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Qualcomm® Dragonwing™ Q-8750 \| 28.044 ms \| 3 - 233 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Snapdragon® 8 Elite Mobile \| 28.044 ms \| 3 - 233 MB \| NPU
	\| Yolo-R \| ONNX \| float \| Snapdragon® 8 Elite Gen 5 Mobile \| 27.585 ms \| 3 - 295 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® X2 Elite \| 19.175 ms \| 2 - 2 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® X Elite \| 30.831 ms \| 41 - 41 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® 8 Gen 3 Mobile \| 20.922 ms \| 1 - 464 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® 8 Gen 1 Mobile \| 42.424 ms \| 3 - 461 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Qualcomm® Dragonwing™ QCS8550 (Proxy) \| 29.41 ms \| 2 - 6 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Qualcomm® QCS8450 \| 42.424 ms \| 3 - 461 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Qualcomm® Dragonwing™ IQ-9075 \| 30.957 ms \| 1 - 6 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Qualcomm® Dragonwing™ IQ-X7181 \| 30.831 ms \| 41 - 41 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Qualcomm® Dragonwing™ Q-8750 \| 17.867 ms \| 2 - 365 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® 8 Elite Mobile \| 17.867 ms \| 2 - 365 MB \| NPU
	\| Yolo-R \| ONNX \| w8a16 \| Snapdragon® 8 Elite Gen 5 Mobile \| 18.762 ms \| 1 - 425 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® X2 Elite \| 8.641 ms \| 2 - 2 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® X Elite \| 19.509 ms \| 2 - 2 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® 8 Gen 3 Mobile \| 12.447 ms \| 2 - 356 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® 8 Gen 1 Mobile \| 28.046 ms \| 2 - 358 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ QCS6490 \| 91.189 ms \| 2 - 7 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ QCS8275 \| 37.802 ms \| 2 - 291 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ QCS8550 (Proxy) \| 18.513 ms \| 2 - 4 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® SA8775P \| 18.744 ms \| 1 - 291 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® SA8650P \| 18.744 ms \| 1 - 291 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® SA8255P \| 18.744 ms \| 1 - 291 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® QCS8450 \| 28.046 ms \| 2 - 358 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ IQ-9075 \| 18.861 ms \| 0 - 4 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ IQ-X7181 \| 19.509 ms \| 2 - 2 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ Q-6690 \| 219.672 ms \| 2 - 399 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ Q-7790 \| 25.965 ms \| 2 - 309 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® Dragonwing™ Q-8750 \| 9.391 ms \| 2 - 304 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® SA7255P \| 37.802 ms \| 2 - 291 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Qualcomm® SA8295P \| 24.277 ms \| 0 - 292 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® 8 Elite Mobile \| 9.391 ms \| 2 - 304 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® 8 Elite Gen 5 Mobile \| 7.859 ms \| 2 - 310 MB \| NPU
	\| Yolo-R \| QNN_DLC \| w8a16 \| Snapdragon® 7 Gen 4 Mobile \| 25.965 ms \| 2 - 309 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)

	## 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).