MobileSam: Optimized for Qualcomm Devices

Transformer based encoder-decoder where prompts specify what to segment in an image thereby allowing segmentation without the need for additional training. The image encoder generates embeddings and the lightweight decoder operates on the embeddings for point and mask based image segmentation.

This is based on the implementation of MobileSam 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 MobileSam 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 MobileSam on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

• Input resolution: 720p (720x1280)
• Model checkpoint: vit_t
• Model size (sam_decoder) (float): 23.7 MB
• Model size (sam_encoder) (float): 26.6 MB
• Number of parameters (sam_decoder): 6.16M
• Number of parameters (sam_encoder): 6.95M

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
decoder	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	2.691 ms	2 - 216 MB	NPU
decoder	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	3.1 ms	0 - 210 MB	NPU
decoder	ONNX	float	Snapdragon® 8 Gen 3 Mobile	4.149 ms	0 - 243 MB	NPU
decoder	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	6.133 ms	4 - 8 MB	NPU
decoder	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	6.784 ms	4 - 7 MB	NPU
decoder	ONNX	float	Qualcomm® Dragonwing™ Q-8750	3.1 ms	0 - 210 MB	NPU
decoder	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	2.384 ms	4 - 193 MB	NPU
decoder	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	2.8 ms	0 - 188 MB	NPU
decoder	QNN_DLC	float	Snapdragon® X2 Elite	2.861 ms	4 - 4 MB	NPU
decoder	QNN_DLC	float	Snapdragon® X Elite	5.821 ms	4 - 4 MB	NPU
decoder	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	3.702 ms	4 - 221 MB	NPU
decoder	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	7.592 ms	0 - 212 MB	NPU
decoder	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	5.665 ms	4 - 11 MB	NPU
decoder	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	5.378 ms	4 - 55 MB	NPU
decoder	QNN_DLC	float	Qualcomm® SA8650P	6.389 ms	0 - 196 MB	NPU
decoder	QNN_DLC	float	Qualcomm® SA8255P	6.389 ms	0 - 196 MB	NPU
decoder	QNN_DLC	float	Qualcomm® QCS8450	7.592 ms	0 - 212 MB	NPU
decoder	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	6.24 ms	4 - 10 MB	NPU
decoder	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	5.821 ms	4 - 4 MB	NPU
decoder	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	2.8 ms	0 - 188 MB	NPU
decoder	QNN_DLC	float	Qualcomm® SA8295P	7.705 ms	0 - 188 MB	NPU
decoder	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	2.238 ms	0 - 182 MB	NPU
decoder	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	2.598 ms	0 - 180 MB	NPU
decoder	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	3.401 ms	0 - 211 MB	NPU
decoder	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	7.238 ms	0 - 206 MB	NPU
decoder	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	5.256 ms	0 - 20 MB	NPU
decoder	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	4.844 ms	0 - 18 MB	NPU
decoder	TFLITE	float	Qualcomm® SA8650P	5.89 ms	0 - 188 MB	NPU
decoder	TFLITE	float	Qualcomm® SA8255P	5.89 ms	0 - 188 MB	NPU
decoder	TFLITE	float	Qualcomm® QCS8450	7.238 ms	0 - 206 MB	NPU
decoder	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	5.766 ms	0 - 19 MB	NPU
decoder	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	2.598 ms	0 - 180 MB	NPU
decoder	TFLITE	float	Qualcomm® SA8295P	7.192 ms	0 - 210 MB	NPU
encoder	ONNX	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	345.95 ms	45 - 949 MB	NPU
encoder	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	383.467 ms	91 - 761 MB	NPU
encoder	ONNX	float	Snapdragon® 8 Gen 3 Mobile	172.741 ms	102 - 1345 MB	NPU
encoder	ONNX	float	Qualcomm® Dragonwing™ IQ-8275	240.237 ms	54 - 58 MB	NPU
encoder	ONNX	float	Qualcomm® Dragonwing™ IQ-9075	217.33 ms	58 - 62 MB	NPU
encoder	ONNX	float	Qualcomm® Dragonwing™ Q-8750	383.467 ms	91 - 761 MB	NPU
encoder	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	639.527 ms	12 - 1438 MB	NPU
encoder	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	687.388 ms	0 - 1296 MB	NPU
encoder	QNN_DLC	float	Snapdragon® X2 Elite	398.212 ms	12 - 12 MB	NPU
encoder	QNN_DLC	float	Snapdragon® X Elite	831.111 ms	12 - 12 MB	NPU
encoder	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	79.943 ms	12 - 1647 MB	NPU
encoder	QNN_DLC	float	Snapdragon® 8 Gen 1 Mobile	203.305 ms	0 - 1332 MB	NPU
encoder	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-8275	987.704 ms	12 - 31 MB	NPU
encoder	QNN_DLC	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	966.327 ms	12 - 836 MB	NPU
encoder	QNN_DLC	float	Qualcomm® SA8650P	100.287 ms	1 - 1549 MB	NPU
encoder	QNN_DLC	float	Qualcomm® SA8255P	100.287 ms	1 - 1549 MB	NPU
encoder	QNN_DLC	float	Qualcomm® QCS8450	203.305 ms	0 - 1332 MB	NPU
encoder	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-9075	915.472 ms	12 - 30 MB	NPU
encoder	QNN_DLC	float	Qualcomm® Dragonwing™ IQ-X7181	831.111 ms	12 - 12 MB	NPU
encoder	QNN_DLC	float	Qualcomm® Dragonwing™ Q-8750	687.388 ms	0 - 1296 MB	NPU
encoder	QNN_DLC	float	Qualcomm® SA8295P	247.754 ms	1 - 1100 MB	NPU
encoder	TFLITE	float	Snapdragon® 8 Elite Gen 5 For Galaxy Mobile	400.866 ms	4 - 1315 MB	NPU
encoder	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	445.047 ms	4 - 1203 MB	NPU
encoder	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	490.256 ms	74 - 1562 MB	NPU
encoder	TFLITE	float	Snapdragon® 8 Gen 1 Mobile	201.436 ms	4 - 1341 MB	NPU
encoder	TFLITE	float	Qualcomm® Dragonwing™ IQ-8275	637.235 ms	0 - 50 MB	NPU
encoder	TFLITE	float	Qualcomm® Dragonwing™ QCS8550 (Proxy)	627.881 ms	0 - 4 MB	NPU
encoder	TFLITE	float	Qualcomm® SA8650P	100.639 ms	0 - 1665 MB	NPU
encoder	TFLITE	float	Qualcomm® SA8255P	100.639 ms	0 - 1665 MB	NPU
encoder	TFLITE	float	Qualcomm® QCS8450	201.436 ms	4 - 1341 MB	NPU
encoder	TFLITE	float	Qualcomm® Dragonwing™ IQ-9075	597.633 ms	0 - 49 MB	NPU
encoder	TFLITE	float	Qualcomm® Dragonwing™ Q-8750	445.047 ms	4 - 1203 MB	NPU
encoder	TFLITE	float	Qualcomm® SA8295P	247.541 ms	0 - 1094 MB	NPU

License

• The license for the original implementation of MobileSam can be found here.

References

• Segment Anything
• 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.