Hugging Face's logo

qualcomm
/
FFNet-122NS-LowRes

Image Segmentation
PyTorch
android
Model card Files
xet
 Community

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 2.45, ONNX Runtime 1.25.0 Download
ONNX w8a8 Universal QAIRT 2.45, ONNX Runtime 1.25.0 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 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® X2 Elite 7.349 ms 207 - 207 MB NPU
FFNet-122NS-LowRes ONNX float Snapdragon® X Elite 14.889 ms 143 - 143 MB NPU
FFNet-122NS-LowRes ONNX float Snapdragon® 8 Gen 3 Mobile 11.185 ms 13 - 253 MB NPU
FFNet-122NS-LowRes ONNX float Snapdragon® 8 Gen 1 Mobile 27.198 ms 11 - 232 MB NPU
FFNet-122NS-LowRes ONNX float Qualcomm® QCS8550 (Proxy) 14.733 ms 0 - 61 MB NPU
FFNet-122NS-LowRes ONNX float Qualcomm® QCS8450 27.198 ms 11 - 232 MB NPU
FFNet-122NS-LowRes ONNX float Snapdragon® 8 Elite Mobile 9.819 ms 6 - 194 MB NPU
FFNet-122NS-LowRes ONNX float Snapdragon® 8 Elite Gen 5 Mobile 5.995 ms 7 - 193 MB NPU
FFNet-122NS-LowRes ONNX float Qualcomm® QCS9075 26.671 ms 6 - 51 MB NPU
FFNet-122NS-LowRes ONNX float Qualcomm® QCS8750 9.819 ms 6 - 194 MB NPU
FFNet-122NS-LowRes ONNX float Qualcomm® QCS7181 14.889 ms 143 - 143 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Snapdragon® X2 Elite 2.017 ms 211 - 211 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Snapdragon® X Elite 4.538 ms 148 - 148 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Snapdragon® 8 Gen 3 Mobile 3.007 ms 0 - 281 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Snapdragon® 8 Gen 1 Mobile 6.076 ms 3 - 286 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Qualcomm® QCS6490 22.196 ms 2 - 47 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Qualcomm® QCS8550 (Proxy) 4.314 ms 0 - 36 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Qualcomm® QCS8450 6.076 ms 3 - 286 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Snapdragon® 8 Elite Gen 5 Mobile 1.764 ms 0 - 223 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Snapdragon® 7 Gen 4 Mobile 9.57 ms 3 - 220 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Qualcomm® QCM6690 77.396 ms 3 - 235 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Qualcomm® QCS9075 4.594 ms 2 - 47 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Snapdragon® 8 Elite Mobile 2.355 ms 0 - 218 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Qualcomm® QCS7790 9.57 ms 3 - 220 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Qualcomm® QCS8750 2.355 ms 0 - 218 MB NPU
FFNet-122NS-LowRes ONNX w8a8 Qualcomm® QCS7181 4.538 ms 148 - 148 MB NPU
FFNet-122NS-LowRes QNN_DLC float Snapdragon® X2 Elite 8.077 ms 6 - 6 MB NPU
FFNet-122NS-LowRes QNN_DLC float Snapdragon® X Elite 17.284 ms 6 - 6 MB NPU
FFNet-122NS-LowRes QNN_DLC float Snapdragon® 8 Gen 3 Mobile 11.52 ms 6 - 242 MB NPU
FFNet-122NS-LowRes QNN_DLC float Snapdragon® 8 Gen 1 Mobile 31.259 ms 6 - 231 MB NPU
FFNet-122NS-LowRes QNN_DLC float Qualcomm® QCS8275 89.255 ms 1 - 183 MB NPU
FFNet-122NS-LowRes QNN_DLC float Qualcomm® QCS8550 (Proxy) 16.551 ms 6 - 8 MB NPU
FFNet-122NS-LowRes QNN_DLC float Qualcomm® QCS8450 31.259 ms 6 - 231 MB NPU
FFNet-122NS-LowRes QNN_DLC float Snapdragon® 8 Elite Mobile 8.365 ms 6 - 193 MB NPU
FFNet-122NS-LowRes QNN_DLC float Qualcomm® SA8295P 27.68 ms 0 - 177 MB NPU
FFNet-122NS-LowRes QNN_DLC float Snapdragon® 8 Elite Gen 5 Mobile 6.164 ms 6 - 205 MB NPU
FFNet-122NS-LowRes QNN_DLC float Qualcomm® SA7255P 89.255 ms 1 - 183 MB NPU
FFNet-122NS-LowRes QNN_DLC float Qualcomm® QCS9075 29.547 ms 6 - 16 MB NPU
FFNet-122NS-LowRes QNN_DLC float Qualcomm® QCS8750 8.365 ms 6 - 193 MB NPU
FFNet-122NS-LowRes QNN_DLC float Qualcomm® QCS7181 17.284 ms 6 - 6 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Snapdragon® X2 Elite 2.763 ms 2 - 2 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Snapdragon® X Elite 6.354 ms 2 - 2 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Snapdragon® 8 Gen 3 Mobile 4.059 ms 2 - 160 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Snapdragon® 8 Gen 1 Mobile 7.077 ms 2 - 160 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Qualcomm® QCS6490 21.529 ms 1 - 5 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Qualcomm® QCS8275 15.7 ms 2 - 98 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Qualcomm® QCS8550 (Proxy) 5.771 ms 2 - 22 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Qualcomm® QCS8450 7.077 ms 2 - 160 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Snapdragon® 8 Elite Gen 5 Mobile 2.167 ms 2 - 214 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Snapdragon® 7 Gen 4 Mobile 10.764 ms 2 - 221 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Qualcomm® QCM6690 78.581 ms 2 - 212 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Qualcomm® QCS9075 7.266 ms 1 - 5 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Qualcomm® SA7255P 15.7 ms 2 - 98 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Snapdragon® 8 Elite Mobile 2.868 ms 2 - 206 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Qualcomm® SA8295P 9.156 ms 1 - 97 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Qualcomm® QCS7790 10.764 ms 2 - 221 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Qualcomm® QCS8750 2.868 ms 2 - 206 MB NPU
FFNet-122NS-LowRes QNN_DLC w8a8 Qualcomm® QCS7181 6.354 ms 2 - 2 MB NPU
FFNet-122NS-LowRes TFLITE float Snapdragon® 8 Gen 3 Mobile 11.563 ms 0 - 313 MB NPU
FFNet-122NS-LowRes TFLITE float Snapdragon® 8 Gen 1 Mobile 31.555 ms 2 - 300 MB NPU
FFNet-122NS-LowRes TFLITE float Qualcomm® QCS8275 89.282 ms 3 - 216 MB NPU
FFNet-122NS-LowRes TFLITE float Qualcomm® QCS8550 (Proxy) 16.556 ms 2 - 5 MB NPU
FFNet-122NS-LowRes TFLITE float Qualcomm® SA8775P 137.239 ms 3 - 26 MB GPU
FFNet-122NS-LowRes TFLITE float Qualcomm® SA8650P 137.239 ms 3 - 26 MB GPU
FFNet-122NS-LowRes TFLITE float Qualcomm® SA8255P 137.239 ms 3 - 26 MB GPU
FFNet-122NS-LowRes TFLITE float Qualcomm® QCS8450 31.555 ms 2 - 300 MB NPU
FFNet-122NS-LowRes TFLITE float Snapdragon® 8 Elite Mobile 8.356 ms 0 - 216 MB NPU
FFNet-122NS-LowRes TFLITE float Qualcomm® SA8295P 27.743 ms 2 - 210 MB NPU
FFNet-122NS-LowRes TFLITE float Snapdragon® 8 Elite Gen 5 Mobile 6.147 ms 2 - 230 MB NPU
FFNet-122NS-LowRes TFLITE float Qualcomm® SA7255P 89.282 ms 3 - 216 MB NPU
FFNet-122NS-LowRes TFLITE float Qualcomm® QCS9075 29.485 ms 0 - 73 MB NPU
FFNet-122NS-LowRes TFLITE float Qualcomm® QCS8750 8.356 ms 0 - 216 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Snapdragon® 8 Gen 3 Mobile 2.766 ms 0 - 159 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Snapdragon® 8 Gen 1 Mobile 5.262 ms 1 - 164 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® QCS6490 16.432 ms 1 - 36 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® QCS8275 12.472 ms 1 - 94 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® QCS8550 (Proxy) 3.828 ms 1 - 4 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® SA8775P 140.549 ms 10 - 38 MB GPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® SA8650P 140.549 ms 10 - 38 MB GPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® SA8255P 140.549 ms 10 - 38 MB GPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® QCS8450 5.262 ms 1 - 164 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Snapdragon® 8 Elite Gen 5 Mobile 1.557 ms 0 - 210 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Snapdragon® 7 Gen 4 Mobile 9.111 ms 0 - 213 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® QCM6690 74.7 ms 1 - 207 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® QCS9075 4.738 ms 1 - 36 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® SA7255P 12.472 ms 1 - 94 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Snapdragon® 8 Elite Mobile 2.094 ms 0 - 203 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® SA8295P 7.045 ms 1 - 92 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® QCS7790 9.111 ms 0 - 213 MB NPU
FFNet-122NS-LowRes TFLITE w8a8 Qualcomm® QCS8750 2.094 ms 0 - 203 MB NPU

License

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

References

Community

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
Image Segmentation
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Paper for qualcomm/FFNet-122NS-LowRes