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 2.37, ONNX Runtime 1.23.0	Download
QNN_DLC	float	Universal	QAIRT 2.42	Download
TFLITE	float	Universal	QAIRT 2.42, TFLite 2.17.0	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® X Elite	134.883 ms	33 - 33 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® 8 Gen 3 Mobile	95.525 ms	0 - 613 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® QCS8550 (Proxy)	131.493 ms	0 - 41 MB	NPU
AOT-GAN	ONNX	float	Qualcomm® QCS9075	235.017 ms	4 - 11 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	73.302 ms	6 - 470 MB	NPU
AOT-GAN	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	49.908 ms	0 - 368 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® X Elite	123.938 ms	4 - 4 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	87.065 ms	0 - 715 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	544.041 ms	1 - 533 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	119.365 ms	4 - 7 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA8775P	161.848 ms	2 - 533 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® QCS9075	214.832 ms	4 - 13 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	215.163 ms	3 - 635 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA7255P	544.041 ms	1 - 533 MB	NPU
AOT-GAN	QNN_DLC	float	Qualcomm® SA8295P	179.326 ms	2 - 471 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	69.435 ms	3 - 565 MB	NPU
AOT-GAN	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	47.503 ms	4 - 473 MB	NPU
AOT-GAN	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	86.747 ms	2 - 750 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® QCS8275 (Proxy)	557.097 ms	3 - 549 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® QCS8550 (Proxy)	122.154 ms	3 - 6 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA8775P	168.373 ms	0 - 547 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® QCS9075	213.878 ms	2 - 45 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® QCS8450 (Proxy)	214.174 ms	3 - 666 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA7255P	557.097 ms	3 - 549 MB	NPU
AOT-GAN	TFLITE	float	Qualcomm® SA8295P	183.927 ms	3 - 488 MB	NPU
AOT-GAN	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	69.074 ms	3 - 580 MB	NPU
AOT-GAN	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	47.494 ms	1 - 490 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.