Hugging Face's logo

qualcomm
/
FFNet-40S

Image Segmentation
PyTorch
real_time
android
Model card Files
xet
 Community

FFNet-40S: Optimized for Qualcomm Devices

FFNet-40S 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-40S 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 1.25.0 Download
ONNX w8a8 Universal QAIRT 2.42, 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-40S 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-40S on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

  • Model checkpoint: ffnet40S_dBBB_cityscapes_state_dict_quarts
  • Input resolution: 2048x1024
  • Number of output classes: 19
  • Number of parameters: 13.9M
  • Model size (float): 53.1 MB
  • Model size (w8a8): 13.5 MB

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
FFNet-40S ONNX float Snapdragon® 8 Elite Gen 5 Mobile 12.366 ms 29 - 254 MB NPU
FFNet-40S ONNX float Snapdragon® X2 Elite 13.431 ms 188 - 188 MB NPU
FFNet-40S ONNX float Snapdragon® X Elite 31.269 ms 157 - 157 MB NPU
FFNet-40S ONNX float Snapdragon® 8 Gen 3 Mobile 22.54 ms 30 - 304 MB NPU
FFNet-40S ONNX float Qualcomm® QCS8550 (Proxy) 32.512 ms 24 - 46 MB NPU
FFNet-40S ONNX float Snapdragon® 8 Elite For Galaxy Mobile 15.575 ms 7 - 208 MB NPU
FFNet-40S ONNX float Qualcomm® QCS9075 48.113 ms 24 - 69 MB NPU
FFNet-40S ONNX float Qualcomm® QCS8750 15.575 ms 7 - 208 MB NPU
FFNet-40S ONNX float Qualcomm® QCS7181 31.269 ms 157 - 157 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® 8 Elite Gen 5 Mobile 6.544 ms 7 - 202 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® X2 Elite 6.892 ms 206 - 206 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® X Elite 10.287 ms 171 - 171 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® 8 Gen 3 Mobile 6.567 ms 7 - 250 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® QCS6490 361.873 ms 206 - 241 MB CPU
FFNet-40S ONNX w8a8 Qualcomm® QCS8550 (Proxy) 9.827 ms 0 - 40 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® 7 Gen 4 Mobile 361.513 ms 163 - 173 MB CPU
FFNet-40S ONNX w8a8 Qualcomm® QCM6690 365.897 ms 161 - 170 MB CPU
FFNet-40S ONNX w8a8 Snapdragon® 8 Elite For Galaxy Mobile 7.715 ms 1 - 195 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® QCS9075 12.815 ms 6 - 51 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® QCS7790 361.513 ms 163 - 173 MB CPU
FFNet-40S ONNX w8a8 Qualcomm® QCS8750 7.715 ms 1 - 195 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® QCS7181 10.287 ms 171 - 171 MB NPU
FFNet-40S QNN_DLC float Snapdragon® 8 Elite Gen 5 Mobile 12.627 ms 24 - 261 MB NPU
FFNet-40S QNN_DLC float Snapdragon® X2 Elite 14.539 ms 24 - 24 MB NPU
FFNet-40S QNN_DLC float Snapdragon® X Elite 37.651 ms 24 - 24 MB NPU
FFNet-40S QNN_DLC float Snapdragon® 8 Gen 3 Mobile 24.884 ms 21 - 290 MB NPU
FFNet-40S QNN_DLC float Qualcomm® QCS8275 135.559 ms 24 - 219 MB NPU
FFNet-40S QNN_DLC float Qualcomm® QCS8550 (Proxy) 36.383 ms 24 - 26 MB NPU
FFNet-40S QNN_DLC float Qualcomm® SA8775P 48.924 ms 24 - 219 MB NPU
FFNet-40S QNN_DLC float Qualcomm® SA8650P 48.924 ms 24 - 219 MB NPU
FFNet-40S QNN_DLC float Qualcomm® SA8255P 48.924 ms 24 - 219 MB NPU
FFNet-40S QNN_DLC float Qualcomm® QCS8450 (Proxy) 64.602 ms 5 - 269 MB NPU
FFNet-40S QNN_DLC float Qualcomm® SA8295P 53.685 ms 24 - 224 MB NPU
FFNet-40S QNN_DLC float Snapdragon® 8 Elite For Galaxy Mobile 18.155 ms 17 - 241 MB NPU
FFNet-40S QNN_DLC float Qualcomm® SA7255P 135.559 ms 24 - 219 MB NPU
FFNet-40S QNN_DLC float Qualcomm® QCS9075 61.747 ms 24 - 52 MB NPU
FFNet-40S QNN_DLC float Qualcomm® QCS8750 18.155 ms 17 - 241 MB NPU
FFNet-40S QNN_DLC float Qualcomm® QCS7181 37.651 ms 24 - 24 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® 8 Elite Gen 5 Mobile 5.238 ms 6 - 230 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® X2 Elite 6.128 ms 6 - 6 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® X Elite 15.967 ms 6 - 6 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® 8 Gen 3 Mobile 10.499 ms 6 - 250 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® QCS6490 66.682 ms 6 - 14 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® QCS8275 32.873 ms 6 - 202 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® QCS8550 (Proxy) 15.101 ms 6 - 146 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® SA8775P 15.678 ms 6 - 203 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® SA8650P 15.678 ms 6 - 203 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® SA8255P 15.678 ms 6 - 203 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® SA7255P 32.873 ms 6 - 202 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® 7 Gen 4 Mobile 19.235 ms 6 - 225 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® SA8295P 20.066 ms 6 - 206 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® QCM6690 127.454 ms 6 - 238 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® 8 Elite For Galaxy Mobile 7.105 ms 6 - 218 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® QCS9075 18.277 ms 6 - 14 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® QCS8450 (Proxy) 21.355 ms 6 - 252 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® QCS7790 19.235 ms 6 - 225 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® QCS8750 7.105 ms 6 - 218 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® QCS7181 15.967 ms 6 - 6 MB NPU
FFNet-40S TFLITE float Snapdragon® 8 Elite Gen 5 Mobile 12.676 ms 2 - 250 MB NPU
FFNet-40S TFLITE float Snapdragon® 8 Gen 3 Mobile 24.932 ms 1 - 304 MB NPU
FFNet-40S TFLITE float Qualcomm® QCS8275 135.705 ms 3 - 211 MB NPU
FFNet-40S TFLITE float Qualcomm® QCS8550 (Proxy) 36.524 ms 3 - 5 MB NPU
FFNet-40S TFLITE float Qualcomm® SA8775P 48.974 ms 3 - 212 MB NPU
FFNet-40S TFLITE float Qualcomm® SA8650P 48.974 ms 3 - 212 MB NPU
FFNet-40S TFLITE float Qualcomm® SA8255P 48.974 ms 3 - 212 MB NPU
FFNet-40S TFLITE float Qualcomm® QCS8450 (Proxy) 64.5 ms 3 - 304 MB NPU
FFNet-40S TFLITE float Qualcomm® SA8295P 53.624 ms 3 - 216 MB NPU
FFNet-40S TFLITE float Snapdragon® 8 Elite For Galaxy Mobile 18.058 ms 0 - 236 MB NPU
FFNet-40S TFLITE float Qualcomm® SA7255P 135.705 ms 3 - 211 MB NPU
FFNet-40S TFLITE float Qualcomm® QCS9075 61.864 ms 0 - 56 MB NPU
FFNet-40S TFLITE float Qualcomm® QCS8750 18.058 ms 0 - 236 MB NPU
FFNet-40S TFLITE w8a8 Snapdragon® 8 Elite Gen 5 Mobile 2.914 ms 1 - 224 MB NPU
FFNet-40S TFLITE w8a8 Snapdragon® 8 Gen 3 Mobile 5.456 ms 0 - 247 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® QCS6490 51.87 ms 1 - 23 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® QCS8275 20.073 ms 1 - 196 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® QCS8550 (Proxy) 7.59 ms 1 - 3 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® SA8775P 8.231 ms 1 - 197 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® SA8650P 8.231 ms 1 - 197 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® SA8255P 8.231 ms 1 - 197 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® SA7255P 20.073 ms 1 - 196 MB NPU
FFNet-40S TFLITE w8a8 Snapdragon® 7 Gen 4 Mobile 11.453 ms 0 - 216 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® SA8295P 11.69 ms 0 - 199 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® QCM6690 102.832 ms 0 - 230 MB NPU
FFNet-40S TFLITE w8a8 Snapdragon® 8 Elite For Galaxy Mobile 3.988 ms 0 - 209 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® QCS9075 9.575 ms 0 - 22 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® QCS8450 (Proxy) 12.35 ms 1 - 249 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® QCS7790 11.453 ms 0 - 216 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® QCS8750 3.988 ms 0 - 209 MB NPU

License

  • The license for the original implementation of FFNet-40S 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-40S