Hugging Face's logo

qualcomm
/
FFNet-78S

Image Segmentation
PyTorch
android
Model card Files
xet
 Community

FFNet-78S: Optimized for Qualcomm Devices

FFNet-78S 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-78S 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 2.50, ONNX Runtime 1.27.1 Download
ONNX w8a8 Universal QAIRT 2.50, ONNX Runtime 1.27.1 Download
QNN_DLC float Universal QAIRT 2.50 Download
QNN_DLC w8a8 Universal QAIRT 2.50 Download
TFLITE float Universal QAIRT 2.50 Download
TFLITE w8a8 Universal QAIRT 2.50 Download

For more device-specific assets and performance metrics, visit FFNet-78S 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-78S on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.semantic_segmentation

Model Stats:

  • Input resolution: 2048x1024
  • Model checkpoint: ffnet78S_dBBB_cityscapes_state_dict_quarts
  • Model size (float): 105 MB
  • Model size (w8a8): 26.7 MB
  • Number of output classes: 19
  • Number of parameters: 27.5M

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
FFNet-78S ONNX float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 16.193 ms 5 - 220 MB NPU
FFNet-78S ONNX float Snapdragon® 8 Elite For Galaxy Mobile 20.798 ms 6 - 198 MB NPU
FFNet-78S ONNX float Snapdragon® X2 Elite 17.565 ms 24 - 24 MB NPU
FFNet-78S ONNX float Snapdragon® X Elite 36.117 ms 31 - 31 MB NPU
FFNet-78S ONNX float Snapdragon® 8 Gen 3 Mobile 25.664 ms 16 - 298 MB NPU
FFNet-78S ONNX float Snapdragon® 8 Gen 1 Mobile 59.8 ms 6 - 278 MB NPU
FFNet-78S ONNX float Qualcomm® Dragonwing™ IQ-8275 57.294 ms 24 - 52 MB NPU
FFNet-78S ONNX float Qualcomm® Dragonwing™ QCS8550 (Proxy) 37.064 ms 24 - 55 MB NPU
FFNet-78S ONNX float Qualcomm® QCS8450 59.8 ms 6 - 278 MB NPU
FFNet-78S ONNX float Qualcomm® Dragonwing™ IQ-9075 54.622 ms 24 - 51 MB NPU
FFNet-78S ONNX float Qualcomm® Dragonwing™ IQ-X7181 36.117 ms 31 - 31 MB NPU
FFNet-78S ONNX float Qualcomm® Dragonwing™ Q-8750 20.798 ms 6 - 198 MB NPU
FFNet-78S ONNX w8a8 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 4.245 ms 2 - 210 MB NPU
FFNet-78S ONNX w8a8 Snapdragon® 8 Elite For Galaxy Mobile 5.732 ms 1 - 208 MB NPU
FFNet-78S ONNX w8a8 Snapdragon® X2 Elite 4.743 ms 7 - 7 MB NPU
FFNet-78S ONNX w8a8 Snapdragon® X Elite 11.876 ms 22 - 22 MB NPU
FFNet-78S ONNX w8a8 Snapdragon® 8 Gen 3 Mobile 7.556 ms 7 - 289 MB NPU
FFNet-78S ONNX w8a8 Snapdragon® 8 Gen 1 Mobile 16.275 ms 7 - 290 MB NPU
FFNet-78S ONNX w8a8 Qualcomm® Dragonwing™ QCS6490 49.494 ms 6 - 9 MB NPU
FFNet-78S ONNX w8a8 Qualcomm® Dragonwing™ IQ-8275 11.357 ms 6 - 10 MB NPU
FFNet-78S ONNX w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 11.369 ms 0 - 211 MB NPU
FFNet-78S ONNX w8a8 Qualcomm® QCS8450 16.275 ms 7 - 290 MB NPU
FFNet-78S ONNX w8a8 Qualcomm® Dragonwing™ IQ-9075 12.837 ms 6 - 9 MB NPU
FFNet-78S ONNX w8a8 Qualcomm® Dragonwing™ IQ-X7181 11.876 ms 22 - 22 MB NPU
FFNet-78S ONNX w8a8 Qualcomm® Dragonwing™ Q-6690 153.169 ms 5 - 229 MB NPU
FFNet-78S ONNX w8a8 Qualcomm® Dragonwing™ Q-7790 16.393 ms 7 - 237 MB NPU
FFNet-78S ONNX w8a8 Qualcomm® Dragonwing™ Q-8750 5.732 ms 1 - 208 MB NPU
FFNet-78S ONNX w8a8 Snapdragon® 7 Gen 4 Mobile 16.393 ms 7 - 237 MB NPU
FFNet-78S QNN_DLC float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 15.353 ms 18 - 246 MB NPU
FFNet-78S QNN_DLC float Snapdragon® 8 Elite For Galaxy Mobile 21.967 ms 16 - 231 MB NPU
FFNet-78S QNN_DLC float Snapdragon® X2 Elite 18.025 ms 24 - 24 MB NPU
FFNet-78S QNN_DLC float Snapdragon® X Elite 43.767 ms 24 - 24 MB NPU
FFNet-78S QNN_DLC float Snapdragon® 8 Gen 3 Mobile 29.609 ms 13 - 298 MB NPU
FFNet-78S QNN_DLC float Snapdragon® 8 Gen 1 Mobile 60.87 ms 25 - 297 MB NPU
FFNet-78S QNN_DLC float Qualcomm® Dragonwing™ IQ-8275 62.875 ms 24 - 53 MB NPU
FFNet-78S QNN_DLC float Qualcomm® Dragonwing™ QCS8550 (Proxy) 42.263 ms 24 - 356 MB NPU
FFNet-78S QNN_DLC float Qualcomm® SA8775P 60.657 ms 24 - 221 MB NPU
FFNet-78S QNN_DLC float Qualcomm® SA8650P 60.657 ms 24 - 221 MB NPU
FFNet-78S QNN_DLC float Qualcomm® SA8255P 60.657 ms 24 - 221 MB NPU
FFNet-78S QNN_DLC float Qualcomm® QCS8450 60.87 ms 25 - 297 MB NPU
FFNet-78S QNN_DLC float Qualcomm® Dragonwing™ IQ-9075 61.875 ms 24 - 53 MB NPU
FFNet-78S QNN_DLC float Qualcomm® Dragonwing™ IQ-X7181 43.767 ms 24 - 24 MB NPU
FFNet-78S QNN_DLC float Qualcomm® Dragonwing™ Q-8750 21.967 ms 16 - 231 MB NPU
FFNet-78S QNN_DLC float Qualcomm® SA7255P 188.698 ms 24 - 220 MB NPU
FFNet-78S QNN_DLC float Qualcomm® SA8295P 66.237 ms 24 - 217 MB NPU
FFNet-78S QNN_DLC w8a8 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 6.047 ms 6 - 240 MB NPU
FFNet-78S QNN_DLC w8a8 Snapdragon® 8 Elite For Galaxy Mobile 8.162 ms 6 - 228 MB NPU
FFNet-78S QNN_DLC w8a8 Snapdragon® X2 Elite 7.142 ms 6 - 6 MB NPU
FFNet-78S QNN_DLC w8a8 Snapdragon® X Elite 17.695 ms 6 - 6 MB NPU
FFNet-78S QNN_DLC w8a8 Snapdragon® 8 Gen 3 Mobile 11.735 ms 6 - 282 MB NPU
FFNet-78S QNN_DLC w8a8 Snapdragon® 8 Gen 1 Mobile 19.779 ms 6 - 281 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® Dragonwing™ QCS6490 52.886 ms 6 - 15 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-8275 16.56 ms 4 - 13 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 16.678 ms 6 - 8 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® SA8775P 17.22 ms 6 - 215 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® SA8650P 17.22 ms 6 - 215 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® SA8255P 17.22 ms 6 - 215 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® QCS8450 19.779 ms 6 - 281 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-9075 17.845 ms 6 - 15 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-X7181 17.695 ms 6 - 6 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-6690 170.239 ms 6 - 220 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-7790 21.583 ms 6 - 233 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-8750 8.162 ms 6 - 228 MB NPU
FFNet-78S QNN_DLC w8a8 Qualcomm® SA8295P 23.247 ms 6 - 218 MB NPU
FFNet-78S QNN_DLC w8a8 Snapdragon® 7 Gen 4 Mobile 21.583 ms 6 - 233 MB NPU
FFNet-78S TFLITE float Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 15.391 ms 2 - 261 MB NPU
FFNet-78S TFLITE float Snapdragon® 8 Elite For Galaxy Mobile 21.801 ms 1 - 244 MB NPU
FFNet-78S TFLITE float Snapdragon® 8 Gen 3 Mobile 29.854 ms 2 - 363 MB NPU
FFNet-78S TFLITE float Snapdragon® 8 Gen 1 Mobile 61.479 ms 3 - 353 MB NPU
FFNet-78S TFLITE float Qualcomm® Dragonwing™ IQ-8275 63.101 ms 0 - 83 MB NPU
FFNet-78S TFLITE float Qualcomm® Dragonwing™ QCS8550 (Proxy) 42.795 ms 3 - 5 MB NPU
FFNet-78S TFLITE float Qualcomm® QCS8450 61.479 ms 3 - 353 MB NPU
FFNet-78S TFLITE float Qualcomm® Dragonwing™ IQ-9075 62.073 ms 0 - 83 MB NPU
FFNet-78S TFLITE float Qualcomm® Dragonwing™ Q-8750 21.801 ms 1 - 244 MB NPU
FFNet-78S TFLITE float Qualcomm® SA7255P 188.779 ms 3 - 231 MB NPU
FFNet-78S TFLITE float Qualcomm® SA8295P 66.263 ms 3 - 227 MB NPU
FFNet-78S TFLITE w8a8 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 3.765 ms 1 - 233 MB NPU
FFNet-78S TFLITE w8a8 Snapdragon® 8 Elite For Galaxy Mobile 4.979 ms 0 - 218 MB NPU
FFNet-78S TFLITE w8a8 Snapdragon® 8 Gen 3 Mobile 6.479 ms 1 - 280 MB NPU
FFNet-78S TFLITE w8a8 Snapdragon® 8 Gen 1 Mobile 11.565 ms 1 - 280 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® Dragonwing™ QCS6490 37.86 ms 1 - 37 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® Dragonwing™ IQ-8275 9.773 ms 1 - 38 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 9.063 ms 1 - 3 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® SA8775P 9.757 ms 1 - 210 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® SA8650P 9.757 ms 1 - 210 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® SA8255P 9.757 ms 1 - 210 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® QCS8450 11.565 ms 1 - 280 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® Dragonwing™ IQ-9075 9.61 ms 0 - 36 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® Dragonwing™ Q-6690 141.393 ms 1 - 213 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® Dragonwing™ Q-7790 14.345 ms 1 - 226 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® Dragonwing™ Q-8750 4.979 ms 0 - 218 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® SA7255P 26.179 ms 1 - 208 MB NPU
FFNet-78S TFLITE w8a8 Qualcomm® SA8295P 14.884 ms 1 - 212 MB NPU
FFNet-78S TFLITE w8a8 Snapdragon® 7 Gen 4 Mobile 14.345 ms 1 - 226 MB NPU

License

  • The license for the original implementation of FFNet-78S 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-78S