FCN-ResNet50: Optimized for Qualcomm Devices

FCN_ResNet50 is a machine learning model that can segment images from the COCO dataset. It uses ResNet50 as a backbone.

This is based on the implementation of FCN-ResNet50 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
ONNX	w8a8	Universal	QAIRT 2.42, ONNX Runtime 1.24.3	Download
QNN_DLC	float	Universal	QAIRT 2.45	Download
QNN_DLC	w8a8	Universal	QAIRT 2.45	Download
TFLITE	float	Universal	QAIRT 2.45	Download
TFLITE	w8a8	Universal	QAIRT 2.45	Download

For more device-specific assets and performance metrics, visit FCN-ResNet50 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 FCN-ResNet50 on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

• Model checkpoint: COCO_WITH_VOC_LABELS_V1
• Input resolution: 224x224
• Number of output classes: 21
• Number of parameters: 33.0M
• Model size (float): 126 MB
• Model size (w8a8): 32.2 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
FCN-ResNet50	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	22.638 ms	4 - 325 MB	NPU
FCN-ResNet50	ONNX	float	Snapdragon® 8 Elite Mobile	27.035 ms	2 - 296 MB	NPU
FCN-ResNet50	ONNX	float	Snapdragon® X2 Elite	23.284 ms	63 - 63 MB	NPU
FCN-ResNet50	ONNX	float	Snapdragon® X Elite	43.182 ms	62 - 62 MB	NPU
FCN-ResNet50	ONNX	float	Snapdragon® X Elite	43.182 ms	62 - 62 MB	NPU
FCN-ResNet50	ONNX	float	Snapdragon® 8 Gen 3 Mobile	32.594 ms	2 - 383 MB	NPU
FCN-ResNet50	ONNX	float	Qualcomm® QCS8550 (Proxy)	43.875 ms	3 - 6 MB	NPU
FCN-ResNet50	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	27.035 ms	2 - 296 MB	NPU
FCN-ResNet50	ONNX	float	Qualcomm® QCS9075	76.009 ms	3 - 9 MB	NPU
FCN-ResNet50	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	7.061 ms	1 - 249 MB	NPU
FCN-ResNet50	ONNX	w8a8	Snapdragon® 8 Elite Mobile	8.572 ms	1 - 203 MB	NPU
FCN-ResNet50	ONNX	w8a8	Snapdragon® X2 Elite	7.085 ms	33 - 33 MB	NPU
FCN-ResNet50	ONNX	w8a8	Snapdragon® X Elite	13.869 ms	32 - 32 MB	NPU
FCN-ResNet50	ONNX	w8a8	Snapdragon® X Elite	13.869 ms	32 - 32 MB	NPU
FCN-ResNet50	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	10.048 ms	0 - 279 MB	NPU
FCN-ResNet50	ONNX	w8a8	Qualcomm® QCS6490	902.852 ms	68 - 111 MB	CPU
FCN-ResNet50	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	13.606 ms	0 - 23 MB	NPU
FCN-ResNet50	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	711.294 ms	89 - 98 MB	CPU
FCN-ResNet50	ONNX	w8a8	Qualcomm® QCM6690	814.28 ms	65 - 74 MB	CPU
FCN-ResNet50	ONNX	w8a8	Qualcomm® QCS9075	15.138 ms	1 - 4 MB	NPU
FCN-ResNet50	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	8.572 ms	1 - 203 MB	NPU
FCN-ResNet50	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	711.294 ms	89 - 98 MB	CPU
FCN-ResNet50	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	21.828 ms	3 - 334 MB	NPU
FCN-ResNet50	QNN_DLC	float	Snapdragon® 8 Elite Mobile	26.474 ms	0 - 318 MB	NPU
FCN-ResNet50	QNN_DLC	float	Snapdragon® X2 Elite	22.857 ms	3 - 3 MB	NPU
FCN-ResNet50	QNN_DLC	float	Snapdragon® X Elite	44.651 ms	3 - 3 MB	NPU
FCN-ResNet50	QNN_DLC	float	Snapdragon® X Elite	44.651 ms	3 - 3 MB	NPU
FCN-ResNet50	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	33.562 ms	1 - 376 MB	NPU
FCN-ResNet50	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	44.834 ms	3 - 6 MB	NPU
FCN-ResNet50	QNN_DLC	float	Qualcomm® SA8775P	72.894 ms	1 - 308 MB	NPU
FCN-ResNet50	QNN_DLC	float	Qualcomm® SA8775P	72.894 ms	1 - 308 MB	NPU
FCN-ResNet50	QNN_DLC	float	Qualcomm® SA8775P	72.894 ms	1 - 308 MB	NPU
FCN-ResNet50	QNN_DLC	float	Qualcomm® SA7255P	273.64 ms	1 - 308 MB	NPU
FCN-ResNet50	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	84.028 ms	0 - 268 MB	NPU
FCN-ResNet50	QNN_DLC	float	Qualcomm® SA8295P	77.78 ms	0 - 217 MB	NPU
FCN-ResNet50	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	26.474 ms	0 - 318 MB	NPU
FCN-ResNet50	QNN_DLC	float	Qualcomm® QCS9075	78.406 ms	3 - 8 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	7.43 ms	1 - 245 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Snapdragon® 8 Elite Mobile	8.937 ms	0 - 202 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Snapdragon® X2 Elite	7.864 ms	1 - 1 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Snapdragon® X Elite	15.189 ms	1 - 1 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Snapdragon® X Elite	15.189 ms	1 - 1 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	11.201 ms	1 - 264 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Qualcomm® QCS6490	93.341 ms	1 - 3 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	14.904 ms	1 - 3 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Qualcomm® SA8775P	14.738 ms	1 - 215 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Qualcomm® SA8775P	14.738 ms	1 - 215 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Qualcomm® SA8775P	14.738 ms	1 - 215 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	26.189 ms	1 - 284 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Qualcomm® QCM6690	404.112 ms	1 - 384 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Qualcomm® QCS9075	17.568 ms	3 - 5 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	8.937 ms	0 - 202 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	24.514 ms	1 - 264 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Qualcomm® SA7255P	39.444 ms	1 - 206 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Qualcomm® SA8295P	21.868 ms	1 - 210 MB	NPU
FCN-ResNet50	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	26.189 ms	1 - 284 MB	NPU
FCN-ResNet50	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	20.99 ms	0 - 368 MB	NPU
FCN-ResNet50	TFLITE	float	Snapdragon® 8 Elite Mobile	27.3 ms	0 - 358 MB	NPU
FCN-ResNet50	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	33.047 ms	0 - 427 MB	NPU
FCN-ResNet50	TFLITE	float	Qualcomm® QCS8550 (Proxy)	44.452 ms	0 - 3 MB	NPU
FCN-ResNet50	TFLITE	float	Qualcomm® SA8775P	72.87 ms	0 - 354 MB	NPU
FCN-ResNet50	TFLITE	float	Qualcomm® SA8775P	72.87 ms	0 - 354 MB	NPU
FCN-ResNet50	TFLITE	float	Qualcomm® SA8775P	72.87 ms	0 - 354 MB	NPU
FCN-ResNet50	TFLITE	float	Qualcomm® SA7255P	273.529 ms	0 - 354 MB	NPU
FCN-ResNet50	TFLITE	float	Qualcomm® QCS8450 (Proxy)	84.205 ms	1 - 321 MB	NPU
FCN-ResNet50	TFLITE	float	Qualcomm® SA8295P	77.772 ms	0 - 268 MB	NPU
FCN-ResNet50	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	27.3 ms	0 - 358 MB	NPU
FCN-ResNet50	TFLITE	float	Qualcomm® QCS9075	78.711 ms	0 - 71 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	7.362 ms	0 - 244 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Snapdragon® 8 Elite Mobile	8.595 ms	0 - 200 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	10.946 ms	0 - 266 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Qualcomm® QCS6490	95.703 ms	0 - 39 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	14.263 ms	0 - 3 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Qualcomm® SA8775P	14.705 ms	0 - 207 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Qualcomm® SA8775P	14.705 ms	0 - 207 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Qualcomm® SA8775P	14.705 ms	0 - 207 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	28.426 ms	0 - 284 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Qualcomm® QCM6690	432.67 ms	0 - 383 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Qualcomm® QCS9075	15.641 ms	0 - 35 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	8.595 ms	0 - 200 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	23.176 ms	0 - 264 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Qualcomm® SA7255P	38.227 ms	0 - 206 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Qualcomm® SA8295P	21.288 ms	0 - 208 MB	NPU
FCN-ResNet50	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	28.426 ms	0 - 284 MB	NPU

License

• The license for the original implementation of FCN-ResNet50 can be found here.

References

• Fully Convolutional Networks for Semantic 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.