Image Segmentation
PyTorch
real_time
android

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 2.51, ONNX Runtime 1.30.0 Download
ONNX w8a8 Universal QAIRT 2.51, ONNX Runtime 1.30.0 Download
QNN_DLC float Universal QAIRT 2.51 Download
QNN_DLC w8a8 Universal QAIRT 2.51 Download
TFLITE float Universal QAIRT 2.51 Download
TFLITE w8a8 Universal QAIRT 2.51 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:

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

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 For Galaxy Mobile 11.325 ms 5 - 209 MB NPU
FFNet-40S ONNX float Snapdragon® 8 Elite For Galaxy Mobile 15.404 ms 6 - 184 MB NPU
FFNet-40S ONNX float Snapdragon® X2 Elite 12.203 ms 24 - 24 MB NPU
FFNet-40S ONNX float Snapdragon® X Elite 28.023 ms 24 - 24 MB NPU
FFNet-40S ONNX float Snapdragon® 8 Gen 3 Mobile 25.141 ms 29 - 285 MB NPU
FFNet-40S ONNX float Snapdragon® 8 Gen 1 Mobile 42.959 ms 30 - 283 MB NPU
FFNet-40S ONNX float Qualcomm® Dragonwing™ IQ-8275 42.092 ms 24 - 28 MB NPU
FFNet-40S ONNX float Qualcomm® Dragonwing™ QCS8550 (Proxy) 28.493 ms 24 - 55 MB NPU
FFNet-40S ONNX float Qualcomm® QCS8450 42.959 ms 30 - 283 MB NPU
FFNet-40S ONNX float Qualcomm® Dragonwing™ IQ-9075 42.198 ms 24 - 28 MB NPU
FFNet-40S ONNX float Qualcomm® Dragonwing™ IQ-X7181 28.023 ms 24 - 24 MB NPU
FFNet-40S ONNX float Qualcomm® Dragonwing™ Q-8750 15.404 ms 6 - 184 MB NPU
FFNet-40S ONNX float Snapdragon® 8 Elite Extreme Gen 6 Mobile 10.563 ms 30 - 212 MB NPU
FFNet-40S ONNX float Snapdragon® 8 Elite Gen 6 Mobile 11.899 ms 30 - 208 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 3.55 ms 1 - 189 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® 8 Elite For Galaxy Mobile 4.92 ms 1 - 186 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® X2 Elite 3.851 ms 7 - 7 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® X Elite 10.239 ms 9 - 9 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® 8 Gen 3 Mobile 7.14 ms 6 - 243 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® 8 Gen 1 Mobile 14.074 ms 7 - 247 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® Dragonwing™ IQ-8275 10.201 ms 6 - 10 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 9.626 ms 0 - 12 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® QCS8450 14.074 ms 7 - 247 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® Dragonwing™ IQ-9075 11.658 ms 6 - 10 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® Dragonwing™ IQ-X7181 10.239 ms 9 - 9 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® Dragonwing™ Q-6690 118.739 ms 5 - 200 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® Dragonwing™ Q-7790 14.613 ms 7 - 206 MB NPU
FFNet-40S ONNX w8a8 Qualcomm® Dragonwing™ Q-8750 4.92 ms 1 - 186 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® 7 Gen 4 Mobile 14.613 ms 7 - 206 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® 8 Elite Extreme Gen 6 Mobile 3.113 ms 5 - 187 MB NPU
FFNet-40S ONNX w8a8 Snapdragon® 8 Elite Gen 6 Mobile 4.587 ms 7 - 183 MB NPU
FFNet-40S QNN_DLC float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 10.668 ms 0 - 218 MB NPU
FFNet-40S QNN_DLC float Snapdragon® 8 Elite For Galaxy Mobile 15.083 ms 23 - 229 MB NPU
FFNet-40S QNN_DLC float Snapdragon® X2 Elite 13.092 ms 24 - 24 MB NPU
FFNet-40S QNN_DLC float Snapdragon® X Elite 30.43 ms 24 - 24 MB NPU
FFNet-40S QNN_DLC float Snapdragon® 8 Gen 3 Mobile 26.224 ms 24 - 288 MB NPU
FFNet-40S QNN_DLC float Snapdragon® 8 Gen 1 Mobile 43.049 ms 0 - 257 MB NPU
FFNet-40S QNN_DLC float Qualcomm® Dragonwing™ IQ-8275 43.451 ms 24 - 53 MB NPU
FFNet-40S QNN_DLC float Qualcomm® Dragonwing™ QCS8550 (Proxy) 29.637 ms 24 - 26 MB NPU
FFNet-40S QNN_DLC float Qualcomm® SA8650P 49.196 ms 24 - 210 MB NPU
FFNet-40S QNN_DLC float Qualcomm® SA8255P 49.196 ms 24 - 210 MB NPU
FFNet-40S QNN_DLC float Qualcomm® QCS8450 43.049 ms 0 - 257 MB NPU
FFNet-40S QNN_DLC float Qualcomm® Dragonwing™ IQ-9075 42.949 ms 24 - 53 MB NPU
FFNet-40S QNN_DLC float Qualcomm® Dragonwing™ IQ-X7181 30.43 ms 24 - 24 MB NPU
FFNet-40S QNN_DLC float Qualcomm® Dragonwing™ Q-8750 15.083 ms 23 - 229 MB NPU
FFNet-40S QNN_DLC float Qualcomm® SA7255P 125.517 ms 24 - 210 MB NPU
FFNet-40S QNN_DLC float Qualcomm® SA8295P 46.221 ms 24 - 213 MB NPU
FFNet-40S QNN_DLC float Snapdragon® 8 Elite Extreme Gen 6 Mobile 9.011 ms 12 - 219 MB NPU
FFNet-40S QNN_DLC float Snapdragon® 8 Elite Gen 6 Mobile 13.193 ms 24 - 209 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 5.342 ms 6 - 218 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® 8 Elite For Galaxy Mobile 7.199 ms 6 - 208 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® X2 Elite 6.248 ms 6 - 6 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® X Elite 16.141 ms 6 - 6 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® 8 Gen 3 Mobile 10.503 ms 6 - 244 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® 8 Gen 1 Mobile 17.889 ms 6 - 244 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® Dragonwing™ QCS6490 49.627 ms 6 - 15 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-8275 15.498 ms 1 - 11 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 15.223 ms 6 - 8 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® QCS8450 17.889 ms 6 - 244 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-9075 16.379 ms 6 - 15 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-X7181 16.141 ms 6 - 6 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-6690 137.193 ms 6 - 196 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-7790 18.789 ms 6 - 208 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-8750 7.199 ms 6 - 208 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® SA7255P 33.367 ms 6 - 194 MB NPU
FFNet-40S QNN_DLC w8a8 Qualcomm® SA8295P 20.13 ms 6 - 198 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® 7 Gen 4 Mobile 18.789 ms 6 - 208 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® 8 Elite Extreme Gen 6 Mobile 4.285 ms 6 - 92 MB NPU
FFNet-40S QNN_DLC w8a8 Snapdragon® 8 Elite Gen 6 Mobile 7.107 ms 6 - 179 MB NPU
FFNet-40S TFLITE float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 10.653 ms 2 - 231 MB NPU
FFNet-40S TFLITE float Snapdragon® 8 Elite For Galaxy Mobile 15.22 ms 2 - 220 MB NPU
FFNet-40S TFLITE float Snapdragon® 8 Gen 3 Mobile 26.163 ms 2 - 298 MB NPU
FFNet-40S TFLITE float Snapdragon® 8 Gen 1 Mobile 42.613 ms 2 - 296 MB NPU
FFNet-40S TFLITE float Qualcomm® Dragonwing™ IQ-8275 42.953 ms 0 - 57 MB NPU
FFNet-40S TFLITE float Qualcomm® Dragonwing™ QCS8550 (Proxy) 29.307 ms 2 - 5 MB NPU
FFNet-40S TFLITE float Qualcomm® QCS8450 42.613 ms 2 - 296 MB NPU
FFNet-40S TFLITE float Qualcomm® Dragonwing™ IQ-9075 42.727 ms 0 - 57 MB NPU
FFNet-40S TFLITE float Qualcomm® Dragonwing™ Q-8750 15.22 ms 2 - 220 MB NPU
FFNet-40S TFLITE float Qualcomm® SA8295P 46.223 ms 3 - 204 MB NPU
FFNet-40S TFLITE float Snapdragon® 8 Elite Extreme Gen 6 Mobile 8.992 ms 2 - 219 MB NPU
FFNet-40S TFLITE float Snapdragon® 8 Elite Gen 6 Mobile 12.932 ms 2 - 199 MB NPU
FFNet-40S TFLITE w8a8 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 3.072 ms 1 - 212 MB NPU
FFNet-40S TFLITE w8a8 Snapdragon® 8 Elite For Galaxy Mobile 3.996 ms 0 - 201 MB NPU
FFNet-40S TFLITE w8a8 Snapdragon® 8 Gen 3 Mobile 5.323 ms 1 - 237 MB NPU
FFNet-40S TFLITE w8a8 Snapdragon® 8 Gen 1 Mobile 9.449 ms 1 - 240 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® Dragonwing™ QCS6490 33.696 ms 1 - 24 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® Dragonwing™ IQ-8275 7.864 ms 1 - 24 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 7.569 ms 1 - 3 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® SA8650P 8.273 ms 1 - 188 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® SA8255P 8.273 ms 1 - 188 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® QCS8450 9.449 ms 1 - 240 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® Dragonwing™ IQ-9075 7.826 ms 1 - 24 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® Dragonwing™ Q-6690 109.324 ms 0 - 188 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® Dragonwing™ Q-7790 11.378 ms 0 - 200 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® Dragonwing™ Q-8750 3.996 ms 0 - 201 MB NPU
FFNet-40S TFLITE w8a8 Qualcomm® SA8295P 11.702 ms 1 - 192 MB NPU
FFNet-40S TFLITE w8a8 Snapdragon® 7 Gen 4 Mobile 11.378 ms 0 - 200 MB NPU
FFNet-40S TFLITE w8a8 Snapdragon® 8 Elite Extreme Gen 6 Mobile 2.315 ms 1 - 85 MB NPU
FFNet-40S TFLITE w8a8 Snapdragon® 8 Elite Gen 6 Mobile 3.191 ms 1 - 172 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
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Paper for qualcomm/FFNet-40S