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 1.25.0	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:

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

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
AOT-GAN	ONNX	float	Snapdragon® X2 Elite	54.549 ms	8 - 8 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® X Elite	138.484 ms	32 - 32 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® 8 Gen 3 Mobile	96.028 ms	1 - 709 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® 8 Gen 1 Mobile	211.078 ms	12 - 613 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	134.235 ms	10 - 14 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® QCS8450	211.078 ms	12 - 613 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	167.038 ms	4 - 11 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	51.012 ms	7 - 511 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® 8 Elite Mobile	75.581 ms	8 - 627 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® Dragonwing™ Q-8750	75.581 ms	8 - 627 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	138.484 ms	32 - 32 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® X2 Elite	50.939 ms	4 - 4 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® X Elite	122.224 ms	4 - 4 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	88.796 ms	4 - 695 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	197.5 ms	4 - 609 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® QCS8275	541.002 ms	1 - 543 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	121.809 ms	4 - 7 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA8775P	161.473 ms	1 - 543 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA8650P	161.473 ms	1 - 543 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA8255P	161.473 ms	1 - 543 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® QCS8450	197.5 ms	4 - 609 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	160.822 ms	4 - 13 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	47.054 ms	3 - 492 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA7255P	541.002 ms	1 - 543 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® 8 Elite Mobile	69.803 ms	3 - 573 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA8295P	178.695 ms	1 - 476 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	69.803 ms	3 - 573 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	122.224 ms	4 - 4 MB	NPU
AOT-GAN	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	88.224 ms	2 - 717 MB	NPU
AOT-GAN	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	198.625 ms	3 - 638 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® QCS8275	541.174 ms	0 - 554 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	121.89 ms	3 - 7 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA8775P	161.455 ms	3 - 557 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA8650P	161.455 ms	3 - 557 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA8255P	161.455 ms	3 - 557 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® QCS8450	198.625 ms	3 - 638 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	161.123 ms	3 - 48 MB	NPU
AOT-GAN	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	47.269 ms	2 - 512 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA7255P	541.174 ms	0 - 554 MB	NPU
AOT-GAN	TFLITE	float	Snapdragon® 8 Elite Mobile	69.36 ms	3 - 591 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA8295P	178.669 ms	3 - 494 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	69.36 ms	3 - 591 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.