FFNet-122NS-LowRes: Optimized for Qualcomm Devices

FFNet-122NS-LowRes is a "fuss-free network" that segments street scene images with per-pixel classes like road, sidewalk, and pedestrian. Trained on the Cityscapes dataset.

This is based on the implementation of FFNet-122NS-LowRes 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 2.43	Download
QNN_DLC	w8a8	Universal	QAIRT 2.43	Download
TFLITE	float	Universal	QAIRT 2.43, TFLite 2.19.1	Download
TFLITE	w8a8	Universal	QAIRT 2.43, TFLite 2.19.1	Download

For more device-specific assets and performance metrics, visit FFNet-122NS-LowRes 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 FFNet-122NS-LowRes on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

• Model checkpoint: ffnet122NS_CCC_cityscapes_state_dict_quarts_pre_down
• Input resolution: 1024x512
• Number of output classes: 19
• Number of parameters: 32.1M
• Model size (float): 123 MB
• Model size (w8a8): 31.3 MB

Performance Summary

Model	Runtime	Precision	Chipset	Inference Time (ms)	Peak Memory Range (MB)	Primary Compute Unit
FFNet-122NS-LowRes	ONNX	float	Snapdragon® 8 Elite Gen 5 Mobile	3.499 ms	2 - 181 MB	NPU
FFNet-122NS-LowRes	ONNX	float	Snapdragon® X2 Elite	3.596 ms	57 - 57 MB	NPU
FFNet-122NS-LowRes	ONNX	float	Snapdragon® X Elite	9.047 ms	56 - 56 MB	NPU
FFNet-122NS-LowRes	ONNX	float	Snapdragon® 8 Gen 3 Mobile	7.949 ms	7 - 226 MB	NPU
FFNet-122NS-LowRes	ONNX	float	Qualcomm® QCS8550 (Proxy)	9.831 ms	0 - 60 MB	NPU
FFNet-122NS-LowRes	ONNX	float	Qualcomm® QCS9075	10.222 ms	6 - 14 MB	NPU
FFNet-122NS-LowRes	ONNX	float	Snapdragon® 8 Elite For Galaxy Mobile	6.395 ms	2 - 180 MB	NPU
FFNet-122NS-LowRes	ONNX	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.322 ms	0 - 87 MB	NPU
FFNet-122NS-LowRes	ONNX	w8a8	Snapdragon® X2 Elite	1.285 ms	30 - 30 MB	NPU
FFNet-122NS-LowRes	ONNX	w8a8	Snapdragon® X Elite	2.795 ms	30 - 30 MB	NPU
FFNet-122NS-LowRes	ONNX	w8a8	Snapdragon® 8 Gen 3 Mobile	1.89 ms	0 - 152 MB	NPU
FFNet-122NS-LowRes	ONNX	w8a8	Qualcomm® QCS6490	91.831 ms	54 - 150 MB	CPU
FFNet-122NS-LowRes	ONNX	w8a8	Qualcomm® QCS8550 (Proxy)	2.611 ms	0 - 40 MB	NPU
FFNet-122NS-LowRes	ONNX	w8a8	Qualcomm® QCS9075	3.134 ms	1 - 4 MB	NPU
FFNet-122NS-LowRes	ONNX	w8a8	Qualcomm® QCM6690	85.005 ms	61 - 74 MB	CPU
FFNet-122NS-LowRes	ONNX	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.464 ms	0 - 92 MB	NPU
FFNet-122NS-LowRes	ONNX	w8a8	Snapdragon® 7 Gen 4 Mobile	84.63 ms	61 - 75 MB	CPU
FFNet-122NS-LowRes	QNN_DLC	float	Snapdragon® 8 Elite Gen 5 Mobile	5.632 ms	6 - 171 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	float	Snapdragon® X2 Elite	6.435 ms	6 - 6 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	float	Snapdragon® X Elite	12.971 ms	6 - 6 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	float	Snapdragon® 8 Gen 3 Mobile	8.768 ms	0 - 200 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	float	Qualcomm® QCS8275 (Proxy)	39.562 ms	1 - 162 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	float	Qualcomm® QCS8550 (Proxy)	12.845 ms	6 - 276 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	float	Qualcomm® SA8775P	16.023 ms	1 - 162 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	float	Qualcomm® QCS9075	16.237 ms	6 - 14 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	float	Qualcomm® QCS8450 (Proxy)	30.018 ms	6 - 195 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	float	Qualcomm® SA7255P	39.562 ms	1 - 162 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	float	Qualcomm® SA8295P	17.701 ms	0 - 158 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	float	Snapdragon® 8 Elite For Galaxy Mobile	6.97 ms	0 - 164 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.8 ms	2 - 80 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Snapdragon® X2 Elite	2.083 ms	2 - 2 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Snapdragon® X Elite	4.84 ms	2 - 2 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Snapdragon® 8 Gen 3 Mobile	3.171 ms	2 - 128 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Qualcomm® QCS6490	14.068 ms	2 - 5 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Qualcomm® QCS8275 (Proxy)	9.257 ms	1 - 76 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Qualcomm® QCS8550 (Proxy)	4.518 ms	2 - 3 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Qualcomm® SA8775P	4.942 ms	2 - 78 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Qualcomm® QCS9075	5.892 ms	1 - 4 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Qualcomm® QCM6690	27.693 ms	1 - 199 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Qualcomm® QCS8450 (Proxy)	6.869 ms	2 - 126 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Qualcomm® SA7255P	9.257 ms	1 - 76 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Qualcomm® SA8295P	6.061 ms	1 - 74 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	2.184 ms	2 - 75 MB	NPU
FFNet-122NS-LowRes	QNN_DLC	w8a8	Snapdragon® 7 Gen 4 Mobile	5.75 ms	0 - 191 MB	NPU
FFNet-122NS-LowRes	TFLITE	float	Snapdragon® 8 Elite Gen 5 Mobile	5.395 ms	1 - 195 MB	NPU
FFNet-122NS-LowRes	TFLITE	float	Snapdragon® 8 Gen 3 Mobile	8.605 ms	1 - 281 MB	NPU
FFNet-122NS-LowRes	TFLITE	float	Qualcomm® QCS8275 (Proxy)	39.925 ms	1 - 193 MB	NPU
FFNet-122NS-LowRes	TFLITE	float	Qualcomm® QCS8550 (Proxy)	12.738 ms	1 - 3 MB	NPU
FFNet-122NS-LowRes	TFLITE	float	Qualcomm® SA8775P	16.016 ms	1 - 193 MB	NPU
FFNet-122NS-LowRes	TFLITE	float	Qualcomm® QCS9075	16.412 ms	0 - 71 MB	NPU
FFNet-122NS-LowRes	TFLITE	float	Qualcomm® QCS8450 (Proxy)	29.929 ms	0 - 269 MB	NPU
FFNet-122NS-LowRes	TFLITE	float	Qualcomm® SA7255P	39.925 ms	1 - 193 MB	NPU
FFNet-122NS-LowRes	TFLITE	float	Qualcomm® SA8295P	18.044 ms	1 - 191 MB	NPU
FFNet-122NS-LowRes	TFLITE	float	Snapdragon® 8 Elite For Galaxy Mobile	6.872 ms	1 - 198 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Snapdragon® 8 Elite Gen 5 Mobile	1.313 ms	0 - 76 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Snapdragon® 8 Gen 3 Mobile	1.863 ms	0 - 134 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Qualcomm® QCS6490	9.031 ms	0 - 35 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Qualcomm® QCS8275 (Proxy)	5.958 ms	0 - 71 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Qualcomm® QCS8550 (Proxy)	2.616 ms	0 - 2 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Qualcomm® SA8775P	3.081 ms	0 - 74 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Qualcomm® QCS9075	3.061 ms	0 - 35 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Qualcomm® QCM6690	20.922 ms	0 - 197 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Qualcomm® QCS8450 (Proxy)	3.268 ms	0 - 128 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Qualcomm® SA7255P	5.958 ms	0 - 71 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Qualcomm® SA8295P	3.861 ms	0 - 69 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Snapdragon® 8 Elite For Galaxy Mobile	1.438 ms	0 - 71 MB	NPU
FFNet-122NS-LowRes	TFLITE	w8a8	Snapdragon® 7 Gen 4 Mobile	3.722 ms	0 - 188 MB	NPU

License

• The license for the original implementation of FFNet-122NS-LowRes can be found here.

References

• Simple and Efficient Architectures 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.