Hugging Face's logo

qualcomm
/
GPUNet

Image Classification
PyTorch
backbone
bu_auto
android
Model card Files
xet
 Community

GPUNet: Optimized for Qualcomm Devices

GPUNet is a machine learning model that can classify images from the Imagenet dataset. It can also be used as a backbone in building more complex models for specific use cases.

This is based on the implementation of GPUNet 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 w8a16 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 w8a16 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 GPUNet 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 GPUNet on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

  • Input resolution: 224x224
  • Model checkpoint: Imagenet
  • Model size (float): 45.28MB
  • Model size (w8a8): 21.3MB
  • Number of parameters: 10.49M

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
GPUNet ONNX float Snapdragon® X2 Elite 0.476 ms 2 - 2 MB NPU
GPUNet ONNX float Snapdragon® X Elite 0.977 ms 25 - 25 MB NPU
GPUNet ONNX float Snapdragon® 8 Gen 3 Mobile 0.763 ms 0 - 61 MB NPU
GPUNet ONNX float Snapdragon® 8 Gen 1 Mobile 1.713 ms 1 - 63 MB NPU
GPUNet ONNX float Qualcomm® Dragonwing™ IQ-8275 1.499 ms 1 - 4 MB NPU
GPUNet ONNX float Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.012 ms 0 - 29 MB NPU
GPUNet ONNX float Qualcomm® QCS8450 1.713 ms 1 - 63 MB NPU
GPUNet ONNX float Qualcomm® Dragonwing™ IQ-9075 1.389 ms 1 - 3 MB NPU
GPUNet ONNX float Qualcomm® Dragonwing™ IQ-X7181 0.977 ms 25 - 25 MB NPU
GPUNet ONNX float Qualcomm® Dragonwing™ Q-8750 0.598 ms 0 - 32 MB NPU
GPUNet ONNX float Snapdragon® 8 Elite Mobile 0.598 ms 0 - 32 MB NPU
GPUNet ONNX float Snapdragon® 8 Elite Gen 5 Mobile 0.518 ms 0 - 33 MB NPU
GPUNet ONNX w8a16 Snapdragon® X2 Elite 0.387 ms 1 - 1 MB NPU
GPUNet ONNX w8a16 Snapdragon® X Elite 0.835 ms 12 - 12 MB NPU
GPUNet ONNX w8a16 Snapdragon® 8 Gen 3 Mobile 0.581 ms 0 - 70 MB NPU
GPUNet ONNX w8a16 Snapdragon® 8 Gen 1 Mobile 1.118 ms 0 - 69 MB NPU
GPUNet ONNX w8a16 Qualcomm® Dragonwing™ QCS6490 3.448 ms 0 - 3 MB NPU
GPUNet ONNX w8a16 Qualcomm® Dragonwing™ IQ-8275 0.874 ms 0 - 4 MB NPU
GPUNet ONNX w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 0.814 ms 0 - 16 MB NPU
GPUNet ONNX w8a16 Qualcomm® QCS8450 1.118 ms 0 - 69 MB NPU
GPUNet ONNX w8a16 Qualcomm® Dragonwing™ IQ-9075 0.93 ms 0 - 3 MB NPU
GPUNet ONNX w8a16 Qualcomm® Dragonwing™ IQ-X7181 0.835 ms 12 - 12 MB NPU
GPUNet ONNX w8a16 Qualcomm® Dragonwing™ Q-6690 5.697 ms 0 - 173 MB NPU
GPUNet ONNX w8a16 Qualcomm® Dragonwing™ Q-7790 1.056 ms 0 - 47 MB NPU
GPUNet ONNX w8a16 Qualcomm® Dragonwing™ Q-8750 0.434 ms 0 - 45 MB NPU
GPUNet ONNX w8a16 Snapdragon® 8 Elite Mobile 0.434 ms 0 - 45 MB NPU
GPUNet ONNX w8a16 Snapdragon® 8 Elite Gen 5 Mobile 0.371 ms 0 - 48 MB NPU
GPUNet ONNX w8a16 Snapdragon® 7 Gen 4 Mobile 1.056 ms 0 - 47 MB NPU
GPUNet ONNX w8a8 Snapdragon® X2 Elite 0.337 ms 1 - 1 MB NPU
GPUNet ONNX w8a8 Snapdragon® X Elite 0.6 ms 12 - 12 MB NPU
GPUNet ONNX w8a8 Snapdragon® 8 Gen 3 Mobile 0.456 ms 0 - 67 MB NPU
GPUNet ONNX w8a8 Snapdragon® 8 Gen 1 Mobile 0.866 ms 0 - 66 MB NPU
GPUNet ONNX w8a8 Qualcomm® Dragonwing™ QCS6490 2.117 ms 0 - 3 MB NPU
GPUNet ONNX w8a8 Qualcomm® Dragonwing™ IQ-8275 0.686 ms 0 - 4 MB NPU
GPUNet ONNX w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 0.62 ms 0 - 17 MB NPU
GPUNet ONNX w8a8 Qualcomm® QCS8450 0.866 ms 0 - 66 MB NPU
GPUNet ONNX w8a8 Qualcomm® Dragonwing™ IQ-9075 0.753 ms 0 - 3 MB NPU
GPUNet ONNX w8a8 Qualcomm® Dragonwing™ IQ-X7181 0.6 ms 12 - 12 MB NPU
GPUNet ONNX w8a8 Qualcomm® Dragonwing™ Q-6690 3.29 ms 0 - 47 MB NPU
GPUNet ONNX w8a8 Qualcomm® Dragonwing™ Q-7790 0.754 ms 0 - 45 MB NPU
GPUNet ONNX w8a8 Qualcomm® Dragonwing™ Q-8750 0.388 ms 0 - 43 MB NPU
GPUNet ONNX w8a8 Snapdragon® 8 Elite Mobile 0.388 ms 0 - 43 MB NPU
GPUNet ONNX w8a8 Snapdragon® 8 Elite Gen 5 Mobile 0.354 ms 0 - 49 MB NPU
GPUNet ONNX w8a8 Snapdragon® 7 Gen 4 Mobile 0.754 ms 0 - 45 MB NPU
GPUNet QNN_DLC float Snapdragon® X2 Elite 0.646 ms 1 - 1 MB NPU
GPUNet QNN_DLC float Snapdragon® X Elite 1.391 ms 1 - 1 MB NPU
GPUNet QNN_DLC float Snapdragon® 8 Gen 3 Mobile 0.909 ms 0 - 56 MB NPU
GPUNet QNN_DLC float Snapdragon® 8 Gen 1 Mobile 2.399 ms 0 - 63 MB NPU
GPUNet QNN_DLC float Qualcomm® Dragonwing™ IQ-8275 1.631 ms 1 - 4 MB NPU
GPUNet QNN_DLC float Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.268 ms 1 - 2 MB NPU
GPUNet QNN_DLC float Qualcomm® SA8775P 1.704 ms 1 - 33 MB NPU
GPUNet QNN_DLC float Qualcomm® SA8650P 1.704 ms 1 - 33 MB NPU
GPUNet QNN_DLC float Qualcomm® SA8255P 1.704 ms 1 - 33 MB NPU
GPUNet QNN_DLC float Qualcomm® QCS8450 2.399 ms 0 - 63 MB NPU
GPUNet QNN_DLC float Qualcomm® Dragonwing™ IQ-9075 1.576 ms 1 - 3 MB NPU
GPUNet QNN_DLC float Qualcomm® Dragonwing™ IQ-X7181 1.391 ms 1 - 1 MB NPU
GPUNet QNN_DLC float Qualcomm® Dragonwing™ Q-8750 0.703 ms 1 - 34 MB NPU
GPUNet QNN_DLC float Qualcomm® SA7255P 4.704 ms 1 - 30 MB NPU
GPUNet QNN_DLC float Qualcomm® SA8295P 2.27 ms 0 - 32 MB NPU
GPUNet QNN_DLC float Snapdragon® 8 Elite Mobile 0.703 ms 1 - 34 MB NPU
GPUNet QNN_DLC float Snapdragon® 8 Elite Gen 5 Mobile 0.573 ms 1 - 35 MB NPU
GPUNet QNN_DLC w8a16 Snapdragon® X2 Elite 0.566 ms 0 - 0 MB NPU
GPUNet QNN_DLC w8a16 Snapdragon® X Elite 1.222 ms 0 - 0 MB NPU
GPUNet QNN_DLC w8a16 Snapdragon® 8 Gen 3 Mobile 0.765 ms 0 - 60 MB NPU
GPUNet QNN_DLC w8a16 Snapdragon® 8 Gen 1 Mobile 1.426 ms 0 - 64 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® Dragonwing™ QCS6490 4.299 ms 2 - 4 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-8275 1.093 ms 0 - 3 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.059 ms 0 - 13 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® SA8775P 1.282 ms 0 - 45 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® SA8650P 1.282 ms 0 - 45 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® SA8255P 1.282 ms 0 - 45 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® QCS8450 1.426 ms 0 - 64 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-9075 1.176 ms 0 - 2 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® Dragonwing™ IQ-X7181 1.222 ms 0 - 0 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-6690 6.548 ms 0 - 168 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-7790 1.296 ms 0 - 44 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® Dragonwing™ Q-8750 0.529 ms 0 - 47 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® SA7255P 2.474 ms 0 - 43 MB NPU
GPUNet QNN_DLC w8a16 Qualcomm® SA8295P 1.685 ms 0 - 41 MB NPU
GPUNet QNN_DLC w8a16 Snapdragon® 8 Elite Mobile 0.529 ms 0 - 47 MB NPU
GPUNet QNN_DLC w8a16 Snapdragon® 8 Elite Gen 5 Mobile 0.441 ms 0 - 43 MB NPU
GPUNet QNN_DLC w8a16 Snapdragon® 7 Gen 4 Mobile 1.296 ms 0 - 44 MB NPU
GPUNet QNN_DLC w8a8 Snapdragon® X2 Elite 0.347 ms 0 - 0 MB NPU
GPUNet QNN_DLC w8a8 Snapdragon® X Elite 0.711 ms 0 - 0 MB NPU
GPUNet QNN_DLC w8a8 Snapdragon® 8 Gen 3 Mobile 0.458 ms 0 - 57 MB NPU
GPUNet QNN_DLC w8a8 Snapdragon® 8 Gen 1 Mobile 0.847 ms 0 - 62 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® Dragonwing™ QCS6490 2.249 ms 2 - 4 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-8275 0.662 ms 0 - 3 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 0.614 ms 0 - 1 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® SA8775P 0.805 ms 0 - 42 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® SA8650P 0.805 ms 0 - 42 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® SA8255P 0.805 ms 0 - 42 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® QCS8450 0.847 ms 0 - 62 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-9075 0.694 ms 0 - 2 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® Dragonwing™ IQ-X7181 0.711 ms 0 - 0 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-6690 3.534 ms 0 - 44 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-7790 0.785 ms 0 - 40 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® Dragonwing™ Q-8750 0.342 ms 0 - 39 MB NPU
GPUNet QNN_DLC w8a8 Qualcomm® SA8295P 1.127 ms 0 - 38 MB NPU
GPUNet QNN_DLC w8a8 Snapdragon® 8 Elite Mobile 0.342 ms 0 - 39 MB NPU
GPUNet QNN_DLC w8a8 Snapdragon® 8 Elite Gen 5 Mobile 0.291 ms 0 - 44 MB NPU
GPUNet QNN_DLC w8a8 Snapdragon® 7 Gen 4 Mobile 0.785 ms 0 - 40 MB NPU
GPUNet TFLITE float Snapdragon® 8 Gen 3 Mobile 0.909 ms 0 - 74 MB NPU
GPUNet TFLITE float Snapdragon® 8 Gen 1 Mobile 2.389 ms 0 - 82 MB NPU
GPUNet TFLITE float Qualcomm® Dragonwing™ IQ-8275 1.629 ms 0 - 28 MB NPU
GPUNet TFLITE float Qualcomm® Dragonwing™ QCS8550 (Proxy) 1.263 ms 0 - 29 MB NPU
GPUNet TFLITE float Qualcomm® SA8775P 1.762 ms 0 - 47 MB NPU
GPUNet TFLITE float Qualcomm® SA8650P 1.762 ms 0 - 47 MB NPU
GPUNet TFLITE float Qualcomm® SA8255P 1.762 ms 0 - 47 MB NPU
GPUNet TFLITE float Qualcomm® QCS8450 2.389 ms 0 - 82 MB NPU
GPUNet TFLITE float Qualcomm® Dragonwing™ IQ-9075 1.608 ms 0 - 27 MB NPU
GPUNet TFLITE float Qualcomm® Dragonwing™ Q-8750 0.698 ms 0 - 44 MB NPU
GPUNet TFLITE float Qualcomm® SA7255P 4.722 ms 0 - 43 MB NPU
GPUNet TFLITE float Qualcomm® SA8295P 2.289 ms 0 - 44 MB NPU
GPUNet TFLITE float Snapdragon® 8 Elite Mobile 0.698 ms 0 - 44 MB NPU
GPUNet TFLITE float Snapdragon® 8 Elite Gen 5 Mobile 0.565 ms 0 - 48 MB NPU
GPUNet TFLITE w8a8 Snapdragon® 8 Gen 3 Mobile 0.342 ms 0 - 55 MB NPU
GPUNet TFLITE w8a8 Snapdragon® 8 Gen 1 Mobile 0.682 ms 0 - 62 MB NPU
GPUNet TFLITE w8a8 Qualcomm® Dragonwing™ QCS6490 1.894 ms 0 - 15 MB NPU
GPUNet TFLITE w8a8 Qualcomm® Dragonwing™ IQ-8275 0.542 ms 0 - 15 MB NPU
GPUNet TFLITE w8a8 Qualcomm® Dragonwing™ QCS8550 (Proxy) 0.428 ms 0 - 2 MB NPU
GPUNet TFLITE w8a8 Qualcomm® SA8775P 0.64 ms 0 - 42 MB NPU
GPUNet TFLITE w8a8 Qualcomm® SA8650P 0.64 ms 0 - 42 MB NPU
GPUNet TFLITE w8a8 Qualcomm® SA8255P 0.64 ms 0 - 42 MB NPU
GPUNet TFLITE w8a8 Qualcomm® QCS8450 0.682 ms 0 - 62 MB NPU
GPUNet TFLITE w8a8 Qualcomm® Dragonwing™ IQ-9075 0.523 ms 0 - 14 MB NPU
GPUNet TFLITE w8a8 Qualcomm® Dragonwing™ Q-6690 3.044 ms 0 - 40 MB NPU
GPUNet TFLITE w8a8 Qualcomm® Dragonwing™ Q-7790 0.615 ms 0 - 39 MB NPU
GPUNet TFLITE w8a8 Qualcomm® Dragonwing™ Q-8750 0.271 ms 0 - 37 MB NPU
GPUNet TFLITE w8a8 Qualcomm® SA7255P 19.083 ms 1 - 25 MB GPU
GPUNet TFLITE w8a8 Qualcomm® SA8295P 0.906 ms 0 - 37 MB NPU
GPUNet TFLITE w8a8 Snapdragon® 8 Elite Mobile 0.271 ms 0 - 37 MB NPU
GPUNet TFLITE w8a8 Snapdragon® 8 Elite Gen 5 Mobile 0.242 ms 0 - 41 MB NPU
GPUNet TFLITE w8a8 Snapdragon® 7 Gen 4 Mobile 0.615 ms 0 - 39 MB NPU

License

  • The license for the original implementation of GPUNet can be found here.

References

Community

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
Image Classification
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Paper for qualcomm/GPUNet