AOT-GAN: Optimized for Qualcomm Devices

AOT-GAN is a machine learning model that allows to erase and in-paint part of given input image.

This is based on the implementation of AOT-GAN 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 1.27.1	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download

For more device-specific assets and performance metrics, visit AOT-GAN 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 AOT-GAN on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_editing

Model Stats:

• Input resolution: 512x512
• Model checkpoint: CelebAHQ
• Model size (float): 58.0 MB
• Number of parameters: 15.2M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
AOT-GAN	ONNX	float	Snapdragon® X2 Elite	54.846 ms	8 - 8 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® X Elite	138.45 ms	32 - 32 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® 8 Gen 3 Mobile	96.214 ms	11 - 723 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® 8 Gen 1 Mobile	208.076 ms	1 - 603 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	188.568 ms	4 - 11 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	135.537 ms	0 - 45 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® QCS8450	208.076 ms	1 - 603 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	167.464 ms	4 - 11 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	138.45 ms	32 - 32 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® Dragonwing™ Q-8750	75.342 ms	7 - 621 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® 8 Elite Mobile	75.342 ms	7 - 621 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	50.917 ms	5 - 510 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® X2 Elite	51.387 ms	4 - 4 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® X Elite	122.325 ms	4 - 4 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	88.451 ms	3 - 695 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	198.219 ms	3 - 606 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	175.03 ms	4 - 13 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	121.888 ms	4 - 6 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA8775P	161.352 ms	1 - 542 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA8650P	161.352 ms	1 - 542 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA8255P	161.352 ms	1 - 542 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® QCS8450	198.219 ms	3 - 606 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	160.953 ms	4 - 13 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	122.325 ms	4 - 4 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	69.645 ms	0 - 571 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA7255P	540.757 ms	1 - 542 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA8295P	178.732 ms	1 - 476 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® 8 Elite Mobile	69.645 ms	0 - 571 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	47.043 ms	3 - 489 MB	NPU
AOT-GAN	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	88.681 ms	2 - 724 MB	NPU
AOT-GAN	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	198.466 ms	3 - 637 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	175.326 ms	3 - 48 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	120.855 ms	3 - 6 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA8775P	161.524 ms	3 - 558 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA8650P	161.524 ms	3 - 558 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA8255P	161.524 ms	3 - 558 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® QCS8450	198.466 ms	3 - 637 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	161.134 ms	3 - 47 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	69.386 ms	2 - 595 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA7255P	541.346 ms	3 - 558 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA8295P	178.686 ms	0 - 491 MB	NPU
AOT-GAN	TFLITE	float	Snapdragon® 8 Elite Mobile	69.386 ms	2 - 595 MB	NPU
AOT-GAN	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	47.223 ms	2 - 512 MB	NPU

License

• The license for the original implementation of AOT-GAN can be found here.

References

• Aggregated Contextual Transformations for High-Resolution Image Inpainting
• Source Model Implementation

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.