LaMa-Dilated: Optimized for Qualcomm Devices

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

This is based on the implementation of LaMa-Dilated 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 2.50, ONNX Runtime 1.30.0	Download
QNN_DLC	float	Universal	QAIRT 2.50	Download
TFLITE	float	Universal	QAIRT 2.50	Download

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

Model Details

Model Type: Model_use_case.image_editing

Model Stats:

• Input resolution: 512x512
• Model checkpoint: Dilated CelebAHQ
• Model size (float): 174 MB
• Number of parameters: 45.6M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
LaMa-Dilated	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	34.491 ms	5 - 381 MB	NPU
LaMa-Dilated	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	43.306 ms	8 - 365 MB	NPU
LaMa-Dilated	ONNX	float	Snapdragon® X2 Elite	35.354 ms	8 - 8 MB	NPU
LaMa-Dilated	ONNX	float	Snapdragon® X Elite	77.363 ms	90 - 90 MB	NPU
LaMa-Dilated	ONNX	float	Snapdragon® 8 Gen 3 Mobile	54.47 ms	1 - 434 MB	NPU
LaMa-Dilated	ONNX	float	Snapdragon® 8 Gen 1 Mobile	108.999 ms	11 - 426 MB	NPU
LaMa-Dilated	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	123.327 ms	6 - 14 MB	NPU
LaMa-Dilated	ONNX	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	78.622 ms	0 - 94 MB	NPU
LaMa-Dilated	ONNX	float	Qualcomm® QCS8450	108.999 ms	11 - 426 MB	NPU
LaMa-Dilated	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	108.002 ms	6 - 14 MB	NPU
LaMa-Dilated	ONNX	float	Qualcomm® Dragonwing™ IQ-X7181	77.363 ms	90 - 90 MB	NPU
LaMa-Dilated	ONNX	float	Qualcomm® Dragonwing™ Q-8750	43.306 ms	8 - 365 MB	NPU
LaMa-Dilated	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	33.853 ms	4 - 360 MB	NPU
LaMa-Dilated	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	41.546 ms	1 - 338 MB	NPU
LaMa-Dilated	QNN_DLC	float	Snapdragon® X2 Elite	35.984 ms	4 - 4 MB	NPU
LaMa-Dilated	QNN_DLC	float	Snapdragon® X Elite	78.352 ms	4 - 4 MB	NPU
LaMa-Dilated	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	53.486 ms	0 - 423 MB	NPU
LaMa-Dilated	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	109.173 ms	4 - 367 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	122.73 ms	4 - 14 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	80.722 ms	4 - 6 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® QCS8450	109.173 ms	4 - 367 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	109.612 ms	4 - 13 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	78.352 ms	4 - 4 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	41.546 ms	1 - 338 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® SA8295P	111.253 ms	1 - 281 MB	NPU
LaMa-Dilated	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	34.082 ms	2 - 373 MB	NPU
LaMa-Dilated	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	41.585 ms	27 - 378 MB	NPU
LaMa-Dilated	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	54.715 ms	3 - 503 MB	NPU
LaMa-Dilated	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	107.523 ms	3 - 443 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	122.849 ms	3 - 106 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	80.815 ms	0 - 49 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® SA8650P	110.59 ms	3 - 354 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® SA8255P	110.59 ms	3 - 354 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® QCS8450	107.523 ms	3 - 443 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	109.907 ms	3 - 106 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	41.585 ms	27 - 378 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® SA8295P	112.568 ms	3 - 293 MB	NPU

License

• The license for the original implementation of LaMa-Dilated can be found here.

References

• Resolution-robust Large Mask Inpainting with Fourier Convolutions
• 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.