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 2.42, ONNX Runtime 1.24.1	Download
ONNX	w8a16	Universal	QAIRT 2.42, ONNX Runtime 1.24.1	Download
QNN_DLC	float	Universal	QAIRT 2.43	Download
QNN_DLC	w8a16	Universal	QAIRT 2.43	Download
TFLITE	float	Universal	QAIRT 2.43, TFLite 2.17.0	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:

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

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 Mobile	34.216 ms	1 - 389 MB	NPU
LaMa-Dilated	ONNX	float	Snapdragon® X2 Elite	36.349 ms	90 - 90 MB	NPU
LaMa-Dilated	ONNX	float	Snapdragon® X Elite	78.189 ms	89 - 89 MB	NPU
LaMa-Dilated	ONNX	float	Snapdragon® 8 Gen 3 Mobile	54.11 ms	0 - 494 MB	NPU
LaMa-Dilated	ONNX	float	Qualcomm® QCS8550 (Proxy)	78.88 ms	0 - 100 MB	NPU
LaMa-Dilated	ONNX	float	Qualcomm® QCS9075	117.624 ms	6 - 13 MB	NPU
LaMa-Dilated	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	41.875 ms	8 - 425 MB	NPU
LaMa-Dilated	ONNX	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	26.682 ms	1 - 536 MB	NPU
LaMa-Dilated	ONNX	w8a16	Snapdragon® X2 Elite	27.939 ms	50 - 50 MB	NPU
LaMa-Dilated	ONNX	w8a16	Snapdragon® X Elite	66.57 ms	50 - 50 MB	NPU
LaMa-Dilated	ONNX	w8a16	Snapdragon® 8 Gen 3 Mobile	45.09 ms	0 - 720 MB	NPU
LaMa-Dilated	ONNX	w8a16	Qualcomm® QCS6490	9928.879 ms	423 - 431 MB	CPU
LaMa-Dilated	ONNX	w8a16	Qualcomm® QCS8550 (Proxy)	64.723 ms	0 - 58 MB	NPU
LaMa-Dilated	ONNX	w8a16	Qualcomm® QCS9075	71.736 ms	3 - 8 MB	NPU
LaMa-Dilated	ONNX	w8a16	Qualcomm® QCM6690	4474.549 ms	366 - 377 MB	CPU
LaMa-Dilated	ONNX	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	32.864 ms	0 - 464 MB	NPU
LaMa-Dilated	ONNX	w8a16	Snapdragon® 7 Gen 4 Mobile	4532.842 ms	360 - 371 MB	CPU
LaMa-Dilated	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	32.073 ms	1 - 386 MB	NPU
LaMa-Dilated	QNN_DLC	float	Snapdragon® X2 Elite	34.311 ms	4 - 4 MB	NPU
LaMa-Dilated	QNN_DLC	float	Snapdragon® X Elite	75.485 ms	4 - 4 MB	NPU
LaMa-Dilated	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	52.786 ms	0 - 466 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	403.644 ms	1 - 360 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	75.548 ms	4 - 6 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® SA8775P	107.116 ms	1 - 360 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® QCS9075	119.657 ms	6 - 15 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	134.587 ms	1 - 416 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® SA7255P	403.644 ms	1 - 360 MB	NPU
LaMa-Dilated	QNN_DLC	float	Qualcomm® SA8295P	113.896 ms	1 - 342 MB	NPU
LaMa-Dilated	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	40.514 ms	1 - 362 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Snapdragon® 8 Elite Gen 5 Mobile	28.004 ms	1 - 764 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Snapdragon® X2 Elite	30.643 ms	2 - 2 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Snapdragon® X Elite	66.657 ms	2 - 2 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Snapdragon® 8 Gen 3 Mobile	44.045 ms	0 - 761 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Qualcomm® QCS8275 (Proxy)	132.75 ms	2 - 613 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Qualcomm® QCS8550 (Proxy)	64.815 ms	2 - 5 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Qualcomm® SA8775P	65.445 ms	2 - 613 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Qualcomm® QCS9075	79.424 ms	1 - 6 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Qualcomm® QCS8450 (Proxy)	115.247 ms	1 - 579 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Qualcomm® SA7255P	132.75 ms	2 - 613 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Qualcomm® SA8295P	90.359 ms	2 - 434 MB	NPU
LaMa-Dilated	QNN_DLC	w8a16	Snapdragon® 8 Elite For Galaxy Mobile	32.598 ms	2 - 615 MB	NPU
LaMa-Dilated	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	32.883 ms	2 - 400 MB	NPU
LaMa-Dilated	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	52.632 ms	0 - 546 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® QCS8275 (Proxy)	403.961 ms	3 - 375 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® QCS8550 (Proxy)	74.802 ms	3 - 132 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® SA8775P	106.247 ms	3 - 374 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® QCS9075	119.515 ms	2 - 104 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® QCS8450 (Proxy)	134.096 ms	3 - 494 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® SA7255P	403.961 ms	3 - 375 MB	NPU
LaMa-Dilated	TFLITE	float	Qualcomm® SA8295P	113.825 ms	3 - 357 MB	NPU
LaMa-Dilated	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	40.345 ms	0 - 378 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.