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 1.25.0 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-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 15.626 ms 189 - 189 MB NPU
FFNet-54S ONNX float Snapdragon® X Elite 34.066 ms 157 - 157 MB NPU
FFNet-54S ONNX float Snapdragon® 8 Gen 3 Mobile 24.403 ms 28 - 314 MB NPU
FFNet-54S ONNX float Snapdragon® 8 Gen 1 Mobile 52.206 ms 29 - 291 MB NPU
FFNet-54S ONNX float Qualcomm® QCS8550 (Proxy) 34.852 ms 24 - 42 MB NPU
FFNet-54S ONNX float Qualcomm® QCS8450 52.206 ms 29 - 291 MB NPU
FFNet-54S ONNX float Snapdragon® 8 Elite Mobile 19.352 ms 6 - 196 MB NPU
FFNet-54S ONNX float Snapdragon® 8 Elite Gen 5 Mobile 14.381 ms 7 - 215 MB NPU
FFNet-54S ONNX float Qualcomm® QCS9075 52.597 ms 24 - 93 MB NPU
FFNet-54S ONNX float Qualcomm® QCS8750 19.352 ms 6 - 196 MB NPU
FFNet-54S ONNX float Qualcomm® QCS7181 34.066 ms 157 - 157 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® X2 Elite 7.394 ms 207 - 207 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® X Elite 11.839 ms 176 - 176 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® 8 Gen 3 Mobile 7.348 ms 7 - 259 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® 8 Gen 1 Mobile 16.818 ms 7 - 260 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® QCS6490 61.09 ms 6 - 51 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® QCS8550 (Proxy) 11.321 ms 0 - 30 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® QCS8450 16.818 ms 7 - 260 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® 8 Elite Gen 5 Mobile 6.969 ms 2 - 201 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® 7 Gen 4 Mobile 14.925 ms 7 - 219 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® QCM6690 127.141 ms 5 - 250 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® QCS9075 13.337 ms 6 - 51 MB NPU
FFNet-54S ONNX w8a8 Snapdragon® 8 Elite Mobile 11.744 ms 0 - 202 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® QCS7790 14.925 ms 7 - 219 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® QCS8750 11.744 ms 0 - 202 MB NPU
FFNet-54S ONNX w8a8 Qualcomm® QCS7181 11.839 ms 176 - 176 MB NPU
FFNet-54S QNN_DLC float Snapdragon® X2 Elite 15.895 ms 24 - 24 MB NPU
FFNet-54S QNN_DLC float Snapdragon® X Elite 39.882 ms 24 - 24 MB NPU
FFNet-54S QNN_DLC float Snapdragon® 8 Gen 3 Mobile 26.862 ms 8 - 288 MB NPU
FFNet-54S QNN_DLC float Snapdragon® 8 Gen 1 Mobile 76.847 ms 24 - 298 MB NPU
FFNet-54S QNN_DLC float Qualcomm® QCS8275 155.46 ms 24 - 225 MB NPU
FFNet-54S QNN_DLC float Qualcomm® QCS8550 (Proxy) 38.615 ms 24 - 26 MB NPU
FFNet-54S QNN_DLC float Qualcomm® QCS8450 76.847 ms 24 - 298 MB NPU
FFNet-54S QNN_DLC float Snapdragon® 8 Elite Mobile 19.564 ms 14 - 235 MB NPU
FFNet-54S QNN_DLC float Qualcomm® SA8295P 59.217 ms 24 - 222 MB NPU
FFNet-54S QNN_DLC float Snapdragon® 8 Elite Gen 5 Mobile 13.611 ms 12 - 257 MB NPU
FFNet-54S QNN_DLC float Qualcomm® SA7255P 155.46 ms 24 - 225 MB NPU
FFNet-54S QNN_DLC float Qualcomm® QCS9075 66.929 ms 24 - 52 MB NPU
FFNet-54S QNN_DLC float Qualcomm® QCS8750 19.564 ms 14 - 235 MB NPU
FFNet-54S QNN_DLC float Qualcomm® QCS7181 39.882 ms 24 - 24 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® X2 Elite 6.477 ms 6 - 6 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® X Elite 16.591 ms 6 - 6 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® 8 Gen 3 Mobile 10.922 ms 6 - 263 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® 8 Gen 1 Mobile 22.201 ms 6 - 268 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® QCS6490 71.901 ms 6 - 14 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® QCS8275 35.581 ms 6 - 211 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® QCS8550 (Proxy) 15.776 ms 6 - 8 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® QCS8450 22.201 ms 6 - 268 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® 8 Elite Gen 5 Mobile 5.539 ms 6 - 240 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® 7 Gen 4 Mobile 20.011 ms 6 - 229 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® QCM6690 139.241 ms 6 - 244 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® QCS9075 18.85 ms 6 - 14 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® SA7255P 35.581 ms 6 - 211 MB NPU
FFNet-54S QNN_DLC w8a8 Snapdragon® 8 Elite Mobile 7.488 ms 2 - 221 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® SA8295P 21.284 ms 6 - 215 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® QCS7790 20.011 ms 6 - 229 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® QCS8750 7.488 ms 2 - 221 MB NPU
FFNet-54S QNN_DLC w8a8 Qualcomm® QCS7181 16.591 ms 6 - 6 MB NPU
FFNet-54S TFLITE float Snapdragon® 8 Gen 3 Mobile 26.756 ms 1 - 332 MB NPU
FFNet-54S TFLITE float Snapdragon® 8 Gen 1 Mobile 76.909 ms 3 - 324 MB NPU
FFNet-54S TFLITE float Qualcomm® QCS8275 155.478 ms 3 - 224 MB NPU
FFNet-54S TFLITE float Qualcomm® QCS8550 (Proxy) 38.253 ms 2 - 5 MB NPU
FFNet-54S TFLITE float Qualcomm® SA8775P 368.763 ms 2 - 31 MB GPU
FFNet-54S TFLITE float Qualcomm® SA8650P 368.763 ms 2 - 31 MB GPU
FFNet-54S TFLITE float Qualcomm® SA8255P 368.763 ms 2 - 31 MB GPU
FFNet-54S TFLITE float Qualcomm® QCS8450 76.909 ms 3 - 324 MB NPU
FFNet-54S TFLITE float Snapdragon® 8 Elite Mobile 19.595 ms 1 - 242 MB NPU
FFNet-54S TFLITE float Qualcomm® SA8295P 59.197 ms 2 - 220 MB NPU
FFNet-54S TFLITE float Snapdragon® 8 Elite Gen 5 Mobile 13.668 ms 1 - 263 MB NPU
FFNet-54S TFLITE float Qualcomm® SA7255P 155.478 ms 3 - 224 MB NPU
FFNet-54S TFLITE float Qualcomm® QCS9075 66.697 ms 0 - 64 MB NPU
FFNet-54S TFLITE float Qualcomm® QCS8750 19.595 ms 1 - 242 MB NPU
FFNet-54S TFLITE w8a8 Snapdragon® 8 Gen 3 Mobile 5.908 ms 1 - 263 MB NPU
FFNet-54S TFLITE w8a8 Snapdragon® 8 Gen 1 Mobile 13.329 ms 1 - 265 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® QCS6490 55.721 ms 1 - 27 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® QCS8275 22.872 ms 1 - 204 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® QCS8550 (Proxy) 8.276 ms 1 - 3 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® SA8775P 318.446 ms 26 - 56 MB GPU
FFNet-54S TFLITE w8a8 Qualcomm® SA8650P 318.446 ms 26 - 56 MB GPU
FFNet-54S TFLITE w8a8 Qualcomm® SA8255P 318.446 ms 26 - 56 MB GPU
FFNet-54S TFLITE w8a8 Qualcomm® QCS8450 13.329 ms 1 - 265 MB NPU
FFNet-54S TFLITE w8a8 Snapdragon® 8 Elite Gen 5 Mobile 3.195 ms 1 - 233 MB NPU
FFNet-54S TFLITE w8a8 Snapdragon® 7 Gen 4 Mobile 12.618 ms 1 - 220 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® QCM6690 116.345 ms 0 - 236 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® QCS9075 10.047 ms 1 - 27 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® SA7255P 22.872 ms 1 - 204 MB NPU
FFNet-54S TFLITE w8a8 Snapdragon® 8 Elite Mobile 4.458 ms 1 - 219 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® SA8295P 13.178 ms 1 - 208 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® QCS7790 12.618 ms 1 - 220 MB NPU
FFNet-54S TFLITE w8a8 Qualcomm® QCS8750 4.458 ms 1 - 219 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