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 1.24.1	Download
QNN_DLC	float	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 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 (SAMEncoder): 6.95M
• Model size (SAMEncoder) (float): 26.6 MB
• Number of parameters (SAMDecoder): 6.16M
• Model size (SAMDecoder) (float): 23.7 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
MobileSAMDecoder	ONNX	float	Snapdragon® X2 Elite	2.953 ms	11 - 11 MB	NPU
MobileSAMDecoder	ONNX	float	Snapdragon® X Elite	7.521 ms	11 - 11 MB	NPU
MobileSAMDecoder	ONNX	float	Snapdragon® 8 Gen 3 Mobile	4.892 ms	4 - 270 MB	NPU
MobileSAMDecoder	ONNX	float	Qualcomm® QCS8550 (Proxy)	6.988 ms	0 - 15 MB	NPU
MobileSAMDecoder	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	3.641 ms	1 - 223 MB	NPU
MobileSAMDecoder	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	2.886 ms	0 - 232 MB	NPU
MobileSAMDecoder	QNN_DLC	float	Snapdragon® X2 Elite	2.609 ms	4 - 4 MB	NPU
MobileSAMDecoder	QNN_DLC	float	Snapdragon® X Elite	5.309 ms	4 - 4 MB	NPU
MobileSAMDecoder	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	3.364 ms	4 - 230 MB	NPU
MobileSAMDecoder	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	11.701 ms	0 - 199 MB	NPU
MobileSAMDecoder	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	4.844 ms	4 - 5 MB	NPU
MobileSAMDecoder	QNN_DLC	float	Qualcomm® SA8775P	5.875 ms	0 - 200 MB	NPU
MobileSAMDecoder	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	8.445 ms	0 - 220 MB	NPU
MobileSAMDecoder	QNN_DLC	float	Qualcomm® SA7255P	11.701 ms	0 - 199 MB	NPU
MobileSAMDecoder	QNN_DLC	float	Qualcomm® SA8295P	7.169 ms	0 - 193 MB	NPU
MobileSAMDecoder	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	2.593 ms	0 - 200 MB	NPU
MobileSAMDecoder	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	2.151 ms	2 - 201 MB	NPU
MobileSAMDecoder	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	3.877 ms	0 - 235 MB	NPU
MobileSAMDecoder	TFLITE	float	Qualcomm® QCS8275 (Proxy)	13.033 ms	0 - 202 MB	NPU
MobileSAMDecoder	TFLITE	float	Qualcomm® QCS8550 (Proxy)	5.605 ms	0 - 18 MB	NPU
MobileSAMDecoder	TFLITE	float	Qualcomm® SA8775P	6.698 ms	0 - 203 MB	NPU
MobileSAMDecoder	TFLITE	float	Qualcomm® QCS8450 (Proxy)	9.541 ms	0 - 225 MB	NPU
MobileSAMDecoder	TFLITE	float	Qualcomm® SA7255P	13.033 ms	0 - 202 MB	NPU
MobileSAMDecoder	TFLITE	float	Qualcomm® SA8295P	8.047 ms	0 - 201 MB	NPU
MobileSAMDecoder	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	2.958 ms	0 - 207 MB	NPU
MobileSAMDecoder	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	2.352 ms	0 - 207 MB	NPU
MobileSAMEncoder	ONNX	float	Snapdragon® X2 Elite	94.358 ms	14 - 14 MB	NPU
MobileSAMEncoder	ONNX	float	Snapdragon® X Elite	173.085 ms	12 - 12 MB	NPU
MobileSAMEncoder	ONNX	float	Snapdragon® 8 Gen 3 Mobile	125.467 ms	20 - 3025 MB	NPU
MobileSAMEncoder	ONNX	float	Qualcomm® QCS8550 (Proxy)	175.267 ms	12 - 32 MB	NPU
MobileSAMEncoder	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	100.697 ms	9 - 1389 MB	NPU
MobileSAMEncoder	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	88.418 ms	20 - 1523 MB	NPU
MobileSAMEncoder	QNN_DLC	float	Snapdragon® X2 Elite	51.746 ms	12 - 12 MB	NPU
MobileSAMEncoder	QNN_DLC	float	Snapdragon® X Elite	100.135 ms	12 - 12 MB	NPU
MobileSAMEncoder	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	66.42 ms	12 - 1635 MB	NPU
MobileSAMEncoder	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	209.122 ms	0 - 1209 MB	NPU
MobileSAMEncoder	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	96.617 ms	12 - 16 MB	NPU
MobileSAMEncoder	QNN_DLC	float	Qualcomm® SA8775P	103.074 ms	0 - 1209 MB	NPU
MobileSAMEncoder	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	470.205 ms	12 - 1731 MB	NPU
MobileSAMEncoder	QNN_DLC	float	Qualcomm® SA7255P	209.122 ms	0 - 1209 MB	NPU
MobileSAMEncoder	QNN_DLC	float	Qualcomm® SA8295P	474.048 ms	0 - 1249 MB	NPU
MobileSAMEncoder	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	49.865 ms	12 - 1217 MB	NPU
MobileSAMEncoder	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	48.392 ms	8 - 1319 MB	NPU
MobileSAMEncoder	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	126.457 ms	0 - 1496 MB	NPU
MobileSAMEncoder	TFLITE	float	Qualcomm® QCS8275 (Proxy)	295.258 ms	4 - 1269 MB	NPU
MobileSAMEncoder	TFLITE	float	Qualcomm® QCS8550 (Proxy)	176.603 ms	0 - 3 MB	NPU
MobileSAMEncoder	TFLITE	float	Qualcomm® SA8775P	179.751 ms	0 - 1265 MB	NPU
MobileSAMEncoder	TFLITE	float	Qualcomm® QCS8450 (Proxy)	590.208 ms	0 - 3175 MB	NPU
MobileSAMEncoder	TFLITE	float	Qualcomm® SA7255P	295.258 ms	4 - 1269 MB	NPU
MobileSAMEncoder	TFLITE	float	Qualcomm® SA8295P	502.789 ms	0 - 1888 MB	NPU
MobileSAMEncoder	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	97.396 ms	4 - 1267 MB	NPU
MobileSAMEncoder	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	78.068 ms	4 - 1348 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.