YOLOv11-Segmentation: Optimized for Mobile Deployment

Real-time object segmentation optimized for mobile and edge by Ultralytics

Ultralytics YOLOv11 is a machine learning model that predicts bounding boxes, segmentation masks and classes of objects in an image.

This model is an implementation of YOLOv11-Segmentation found here.

More details on model performance across various devices, can be found here.

Model Details

• Model Type: Semantic segmentation
• Model Stats:
  • Model checkpoint: YOLO11N-Seg
  • Input resolution: 640x640
  • Number of parameters: 2.9M
  • Model size: 11.1 MB
  • Number of output classes: 80

Model	Device	Chipset	Target Runtime	Inference Time (ms)	Peak Memory Range (MB)	Precision	Primary Compute Unit	Target Model
YOLOv11-Segmentation	Samsung Galaxy S23	Snapdragon® 8 Gen 2	TFLITE	6.633 ms	4 - 30 MB	FP16	NPU	--
YOLOv11-Segmentation	Samsung Galaxy S23	Snapdragon® 8 Gen 2	ONNX	104.404 ms	95 - 108 MB	FP32	CPU	--
YOLOv11-Segmentation	Samsung Galaxy S24	Snapdragon® 8 Gen 3	TFLITE	4.766 ms	0 - 53 MB	FP16	NPU	--
YOLOv11-Segmentation	Samsung Galaxy S24	Snapdragon® 8 Gen 3	ONNX	86.584 ms	97 - 120 MB	FP32	CPU	--
YOLOv11-Segmentation	Snapdragon 8 Elite QRD	Snapdragon® 8 Elite	TFLITE	3.908 ms	0 - 54 MB	FP16	NPU	--
YOLOv11-Segmentation	Snapdragon 8 Elite QRD	Snapdragon® 8 Elite	ONNX	88.869 ms	98 - 114 MB	FP32	CPU	--
YOLOv11-Segmentation	SA7255P ADP	SA7255P	TFLITE	81.11 ms	4 - 50 MB	FP16	NPU	--
YOLOv11-Segmentation	SA8255 (Proxy)	SA8255P Proxy	TFLITE	6.575 ms	4 - 22 MB	FP16	NPU	--
YOLOv11-Segmentation	SA8295P ADP	SA8295P	TFLITE	11.829 ms	4 - 38 MB	FP16	NPU	--
YOLOv11-Segmentation	SA8650 (Proxy)	SA8650P Proxy	TFLITE	6.586 ms	4 - 31 MB	FP16	NPU	--
YOLOv11-Segmentation	SA8775P ADP	SA8775P	TFLITE	9.978 ms	4 - 51 MB	FP16	NPU	--
YOLOv11-Segmentation	QCS8275 (Proxy)	QCS8275 Proxy	TFLITE	81.11 ms	4 - 50 MB	FP16	NPU	--
YOLOv11-Segmentation	QCS8550 (Proxy)	QCS8550 Proxy	TFLITE	6.612 ms	4 - 31 MB	FP16	NPU	--
YOLOv11-Segmentation	QCS9075 (Proxy)	QCS9075 Proxy	TFLITE	9.978 ms	4 - 51 MB	FP16	NPU	--
YOLOv11-Segmentation	QCS8450 (Proxy)	QCS8450 Proxy	TFLITE	10.242 ms	4 - 43 MB	FP16	NPU	--
YOLOv11-Segmentation	Snapdragon X Elite CRD	Snapdragon® X Elite	ONNX	31.289 ms	116 - 116 MB	FP32	CPU	--

License

• The license for the original implementation of YOLOv11-Segmentation can be found here.
• The license for the compiled assets for on-device deployment can be found here

References

• Ultralytics YOLOv11 Docs: Instance Segmentation
• 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.

Usage and Limitations

Model may not be used for or in connection with any of the following applications:

• Accessing essential private and public services and benefits;
• Administration of justice and democratic processes;
• Assessing or recognizing the emotional state of a person;
• Biometric and biometrics-based systems, including categorization of persons based on sensitive characteristics;
• Education and vocational training;
• Employment and workers management;
• Exploitation of the vulnerabilities of persons resulting in harmful behavior;
• General purpose social scoring;
• Law enforcement;
• Management and operation of critical infrastructure;
• Migration, asylum and border control management;
• Predictive policing;
• Real-time remote biometric identification in public spaces;
• Recommender systems of social media platforms;
• Scraping of facial images (from the internet or otherwise); and/or
• Subliminal manipulation