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 2.42, ONNX Runtime 1.24.3	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 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:

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

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
decoder	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	2.952 ms	0 - 242 MB	NPU
decoder	ONNX	float	Snapdragon® X2 Elite	3.025 ms	11 - 11 MB	NPU
decoder	ONNX	float	Snapdragon® X Elite	7.599 ms	11 - 11 MB	NPU
decoder	ONNX	float	Snapdragon® 8 Gen 3 Mobile	4.963 ms	5 - 268 MB	NPU
decoder	ONNX	float	Qualcomm® QCS8550 (Proxy)	7.126 ms	0 - 17 MB	NPU
decoder	ONNX	float	Qualcomm® QCS9075	8.461 ms	4 - 7 MB	NPU
decoder	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	3.676 ms	1 - 231 MB	NPU
decoder	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	2.118 ms	2 - 188 MB	NPU
decoder	QNN_DLC	float	Snapdragon® X2 Elite	2.59 ms	4 - 4 MB	NPU
decoder	QNN_DLC	float	Snapdragon® X Elite	5.173 ms	4 - 4 MB	NPU
decoder	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	3.366 ms	4 - 213 MB	NPU
decoder	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	11.694 ms	1 - 184 MB	NPU
decoder	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	4.808 ms	4 - 6 MB	NPU
decoder	QNN_DLC	float	Qualcomm® SA8775P	5.867 ms	2 - 187 MB	NPU
decoder	QNN_DLC	float	Qualcomm® QCS9075	5.883 ms	4 - 10 MB	NPU
decoder	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	10.521 ms	0 - 212 MB	NPU
decoder	QNN_DLC	float	Qualcomm® SA7255P	11.694 ms	1 - 184 MB	NPU
decoder	QNN_DLC	float	Qualcomm® SA8295P	7.321 ms	0 - 188 MB	NPU
decoder	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	2.58 ms	0 - 183 MB	NPU
decoder	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	2.608 ms	0 - 190 MB	NPU
decoder	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	3.757 ms	0 - 240 MB	NPU
decoder	TFLITE	float	Qualcomm® QCS8275 (Proxy)	12.867 ms	0 - 215 MB	NPU
decoder	TFLITE	float	Qualcomm® QCS8550 (Proxy)	5.613 ms	0 - 8 MB	NPU
decoder	TFLITE	float	Qualcomm® SA8775P	6.617 ms	0 - 215 MB	NPU
decoder	TFLITE	float	Qualcomm® QCS9075	6.868 ms	0 - 19 MB	NPU
decoder	TFLITE	float	Qualcomm® QCS8450 (Proxy)	9.292 ms	0 - 217 MB	NPU
decoder	TFLITE	float	Qualcomm® SA7255P	12.867 ms	0 - 215 MB	NPU
decoder	TFLITE	float	Qualcomm® SA8295P	8.817 ms	0 - 190 MB	NPU
decoder	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	2.992 ms	0 - 186 MB	NPU
encoder	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	87.98 ms	20 - 1523 MB	NPU
encoder	ONNX	float	Snapdragon® X2 Elite	93.865 ms	14 - 14 MB	NPU
encoder	ONNX	float	Snapdragon® X Elite	173.026 ms	12 - 12 MB	NPU
encoder	ONNX	float	Snapdragon® 8 Gen 3 Mobile	125.003 ms	20 - 3038 MB	NPU
encoder	ONNX	float	Qualcomm® QCS8550 (Proxy)	179.446 ms	12 - 38 MB	NPU
encoder	ONNX	float	Qualcomm® QCS9075	201.765 ms	12 - 15 MB	NPU
encoder	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	101.134 ms	9 - 1395 MB	NPU
encoder	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	45.191 ms	10 - 1310 MB	NPU
encoder	QNN_DLC	float	Snapdragon® X2 Elite	48.308 ms	12 - 12 MB	NPU
encoder	QNN_DLC	float	Snapdragon® X Elite	98.3 ms	12 - 12 MB	NPU
encoder	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	66.639 ms	11 - 1557 MB	NPU
encoder	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	209.871 ms	1 - 1193 MB	NPU
encoder	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	95.037 ms	5 - 8 MB	NPU
encoder	QNN_DLC	float	Qualcomm® SA8775P	101.107 ms	0 - 1601 MB	NPU
encoder	QNN_DLC	float	Qualcomm® QCS9075	121.551 ms	12 - 30 MB	NPU
encoder	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	316.36 ms	12 - 1581 MB	NPU
encoder	QNN_DLC	float	Qualcomm® SA7255P	209.871 ms	1 - 1193 MB	NPU
encoder	QNN_DLC	float	Qualcomm® SA8295P	261.882 ms	0 - 1211 MB	NPU
encoder	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	49.263 ms	12 - 1211 MB	NPU
encoder	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	106.887 ms	32 - 1084 MB	NPU
encoder	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	170.886 ms	33 - 2317 MB	NPU
encoder	TFLITE	float	Qualcomm® QCS8275 (Proxy)	363.051 ms	33 - 1119 MB	NPU
encoder	TFLITE	float	Qualcomm® QCS8550 (Proxy)	222.429 ms	33 - 37 MB	NPU
encoder	TFLITE	float	Qualcomm® SA8775P	243.033 ms	33 - 1114 MB	NPU
encoder	TFLITE	float	Qualcomm® QCS9075	313.198 ms	32 - 109 MB	NPU
encoder	TFLITE	float	Qualcomm® QCS8450 (Proxy)	536.317 ms	0 - 1932 MB	NPU
encoder	TFLITE	float	Qualcomm® SA7255P	363.051 ms	33 - 1119 MB	NPU
encoder	TFLITE	float	Qualcomm® SA8295P	373.763 ms	19 - 1147 MB	NPU
encoder	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	129.761 ms	0 - 1074 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.