Hugging Face's logo

qualcomm
/
FFNet-54S

Image Segmentation
PyTorch
android
Model card Files
xet
 Community

FFNet-54S: Optimized for Qualcomm Devices

FFNet-54S 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-54S 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 1.27.1 Download
ONNX w8a8 Universal QAIRT 2.45, ONNX Runtime 1.27.1 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-54S 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-54S on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

  • Model checkpoint: ffnet54S_dBBB_cityscapes_state_dict_quarts
  • Input resolution: 2048x1024
  • Number of output classes: 19
  • Number of parameters: 18.0M
  • Model size (float): 68.8 MB
  • Model size (w8a8): 17.5 MB

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
FFNet-54S ONNX float Snapdragon® X2 Elite 15.619 ms 24 - 24 MB NPU
FFNet-54S ONNX float Snapdragon® X Elite 33.992 ms 24 - 24 MB NPU
FFNet-54S ONNX float Snapdragon® 8 Gen 3 Mobile 24.188 ms 1 - 298 MB NPU
FFNet-54S ONNX float Snapdragon® 8 Gen 1 Mobile 53.1 ms 1 - 269 MB NPU
FFNet-54S ONNX float Qualcomm® Dragonwing™ QCS8550 (Proxy) 34.533 ms 24 - 45 MB NPU
FFNet-54S ONNX float Qualcomm® QCS8450 53.1 ms 1 - 269 MB NPU
FFNet-54S ONNX float Qualcomm® Dragonwing™ IQ-9075 49.489 ms 24 - 51 MB NPU
FFNet-54S ONNX float Qualcomm® Dragonwing™ IQ-X7181 33.992 ms 24 - 24 MB NPU
FFNet-54S ONNX float Qualcomm® Dragonwing™ Q-8750 19.48 ms 5 - 196 MB NPU
FFNet-54S ONNX float Snapdragon® 8 Elite Mobile 19.48 ms 5 - 196 MB NPU
FFNet-54S ONNX float Snapdragon® 8 Elite Gen 5 Mobile 14.557 ms 5 - 218 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® X2 Elite 7.38 ms 7 - 7 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® X Elite 11.92 ms 12 - 12 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® 8 Gen 3 Mobile 7.364 ms 7 - 272 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® 8 Gen 1 Mobile 16.896 ms 7 - 271 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® Dragonwing™ QCS6490 52.585 ms 6 - 9 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 11.315 ms 0 - 16 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® QCS8450 16.896 ms 7 - 271 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® Dragonwing™ IQ-9075 11.671 ms 6 - 9 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® Dragonwing™ IQ-X7181 11.92 ms 12 - 12 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® Dragonwing™ Q-8750 11.79 ms 1 - 203 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® 8 Elite Mobile 11.79 ms 1 - 203 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® 8 Elite Gen 5 Mobile 6.957 ms 1 - 204 MB NPU
FFNet-54S QNN_DLC float Snapdragon® X2 Elite 15.955 ms 24 - 24 MB NPU
FFNet-54S QNN_DLC float Snapdragon® X Elite 39.843 ms 24 - 24 MB NPU
FFNet-54S QNN_DLC float Snapdragon® 8 Gen 3 Mobile 26.766 ms 14 - 293 MB NPU
FFNet-54S QNN_DLC float Snapdragon® 8 Gen 1 Mobile 76.915 ms 24 - 298 MB NPU
FFNet-54S QNN_DLC float Qualcomm® Dragonwing™ QCS8550 (Proxy) 38.439 ms 24 - 26 MB NPU
FFNet-54S QNN_DLC float Qualcomm® SA8775P 53.685 ms 24 - 225 MB NPU
FFNet-54S QNN_DLC float Qualcomm® SA8650P 53.685 ms 24 - 225 MB NPU
FFNet-54S QNN_DLC float Qualcomm® SA8255P 53.685 ms 24 - 225 MB NPU
FFNet-54S QNN_DLC float Qualcomm® QCS8450 76.915 ms 24 - 298 MB NPU
FFNet-54S QNN_DLC float Qualcomm® Dragonwing™ IQ-9075 54.512 ms 24 - 52 MB NPU
FFNet-54S QNN_DLC float Qualcomm® Dragonwing™ IQ-X7181 39.843 ms 24 - 24 MB NPU
FFNet-54S QNN_DLC float Qualcomm® Dragonwing™ Q-8750 19.558 ms 19 - 240 MB NPU
FFNet-54S QNN_DLC float Qualcomm® SA8295P 59.25 ms 24 - 223 MB NPU
FFNet-54S QNN_DLC float Snapdragon® 8 Elite Mobile 19.558 ms 19 - 240 MB NPU
FFNet-54S QNN_DLC float Snapdragon® 8 Elite Gen 5 Mobile 13.765 ms 16 - 260 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® X2 Elite 6.447 ms 6 - 6 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® X Elite 16.671 ms 6 - 6 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® 8 Gen 3 Mobile 10.945 ms 6 - 265 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® 8 Gen 1 Mobile 22.207 ms 6 - 267 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® Dragonwing™ QCS6490 58.925 ms 1 - 9 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 15.741 ms 6 - 8 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® SA8775P 16.269 ms 6 - 212 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® SA8650P 16.269 ms 6 - 212 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® SA8255P 16.269 ms 6 - 212 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® QCS8450 22.207 ms 6 - 267 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-9075 16.373 ms 8 - 16 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-X7181 16.671 ms 6 - 6 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-6690 142.694 ms 6 - 244 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-7790 20.451 ms 6 - 229 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-8750 7.466 ms 6 - 224 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® SA8295P 21.317 ms 6 - 215 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® 8 Elite Mobile 7.466 ms 6 - 224 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® 8 Elite Gen 5 Mobile 5.565 ms 6 - 241 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® 7 Gen 4 Mobile 20.451 ms 6 - 229 MB NPU
FFNet-54S TFLITE float Snapdragon® 8 Gen 3 Mobile 26.863 ms 1 - 332 MB NPU
FFNet-54S TFLITE float Snapdragon® 8 Gen 1 Mobile 77.197 ms 3 - 324 MB NPU
FFNet-54S TFLITE float Qualcomm® Dragonwing™ QCS8550 (Proxy) 38.39 ms 2 - 5 MB NPU
FFNet-54S TFLITE float Qualcomm® SA8775P 53.744 ms 3 - 224 MB NPU
FFNet-54S TFLITE float Qualcomm® SA8650P 53.744 ms 3 - 224 MB NPU
FFNet-54S TFLITE float Qualcomm® SA8255P 53.744 ms 3 - 224 MB NPU
FFNet-54S TFLITE float Qualcomm® QCS8450 77.197 ms 3 - 324 MB NPU
FFNet-54S TFLITE float Qualcomm® Dragonwing™ IQ-9075 54.763 ms 0 - 64 MB NPU
FFNet-54S TFLITE float Qualcomm® Dragonwing™ Q-8750 19.615 ms 0 - 239 MB NPU
FFNet-54S TFLITE float Qualcomm® SA8295P 59.176 ms 2 - 219 MB NPU
FFNet-54S TFLITE float Snapdragon® 8 Elite Mobile 19.615 ms 0 - 239 MB NPU
FFNet-54S TFLITE float Snapdragon® 8 Elite Gen 5 Mobile 13.797 ms 2 - 264 MB NPU
FFNet-54S TFLITE w8a8 Snapdragon® 8 Gen 3 Mobile 5.912 ms 1 - 260 MB NPU
FFNet-54S TFLITE w8a8 Snapdragon® 8 Gen 1 Mobile 13.251 ms 1 - 264 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® Dragonwing™ QCS8275 22.86 ms 1 - 204 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 8.275 ms 1 - 173 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® SA8775P 8.86 ms 1 - 205 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® SA8650P 8.86 ms 1 - 205 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® SA8255P 8.86 ms 1 - 205 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® QCS8450 13.251 ms 1 - 264 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® Dragonwing™ IQ-9075 8.598 ms 0 - 27 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® Dragonwing™ Q-6690 116.752 ms 2 - 238 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® Dragonwing™ Q-7790 12.612 ms 1 - 220 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® Dragonwing™ Q-8750 4.465 ms 0 - 218 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® SA7255P 22.86 ms 1 - 204 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® SA8295P 13.184 ms 0 - 206 MB NPU
FFNet-54S TFLITE w8a8 Snapdragon® 8 Elite Mobile 4.465 ms 0 - 218 MB NPU
FFNet-54S TFLITE w8a8 Snapdragon® 8 Elite Gen 5 Mobile 3.236 ms 1 - 234 MB NPU
FFNet-54S TFLITE w8a8 Snapdragon® 7 Gen 4 Mobile 12.612 ms 1 - 220 MB NPU

License

  • The license for the original implementation of FFNet-54S 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-54S