Hugging Face's logo

qualcomm
/
LeViT

Image Classification
PyTorch
bu_auto
android
Model card Files
xet
 Community

LeViT: Optimized for Qualcomm Devices

LeViT is a vision transformer model that can classify images from the Imagenet dataset.

This is based on the implementation of LeViT 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 w8a16 Universal QAIRT 2.42, ONNX Runtime 1.25.0 Download
ONNX w8a16_mixed_int16 Universal QAIRT 2.42, ONNX Runtime 1.25.0 Download
QNN_DLC float Universal QAIRT 2.45 Download
QNN_DLC w8a16 Universal QAIRT 2.45 Download
QNN_DLC w8a16_mixed_int16 Universal QAIRT 2.45 Download
TFLITE float Universal QAIRT 2.45 Download

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

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

  • Model checkpoint: LeViT-128S
  • Input resolution: 224x224
  • Number of parameters: 7.82M
  • Model size (float): 29.9 MB
  • Model size (w8a16): 8.83 MB

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
LeViT ONNX float Snapdragon® 8 Elite Gen 5 Mobile 0.653 ms 1 - 61 MB NPU
LeViT ONNX float Snapdragon® X2 Elite 0.7 ms 180 - 180 MB NPU
LeViT ONNX float Snapdragon® X Elite 1.314 ms 148 - 148 MB NPU
LeViT ONNX float Snapdragon® 8 Gen 3 Mobile 0.867 ms 1 - 91 MB NPU
LeViT ONNX float Qualcomm® QCS8550 (Proxy) 1.258 ms 0 - 141 MB NPU
LeViT ONNX float Snapdragon® 8 Elite For Galaxy Mobile 0.705 ms 0 - 58 MB NPU
LeViT ONNX float Qualcomm® QCS9075 1.629 ms 1 - 46 MB NPU
LeViT ONNX float Qualcomm® QCS8750 0.705 ms 0 - 58 MB NPU
LeViT ONNX float Qualcomm® QCS7181 1.314 ms 148 - 148 MB NPU
LeViT ONNX w8a16 Snapdragon® 8 Elite Gen 5 Mobile 0.46 ms 0 - 60 MB NPU
LeViT ONNX w8a16 Snapdragon® X2 Elite 0.483 ms 212 - 212 MB NPU
LeViT ONNX w8a16 Snapdragon® X Elite 1.053 ms 149 - 149 MB NPU
LeViT ONNX w8a16 Snapdragon® 8 Gen 3 Mobile 0.696 ms 0 - 87 MB NPU
LeViT ONNX w8a16 Qualcomm® QCS6490 61.639 ms 12 - 16 MB CPU
LeViT ONNX w8a16 Qualcomm® QCS8550 (Proxy) 1.03 ms 0 - 136 MB NPU
LeViT ONNX w8a16 Qualcomm® QCM6690 31.899 ms 13 - 23 MB CPU
LeViT ONNX w8a16 Snapdragon® 7 Gen 4 Mobile 28.139 ms 13 - 23 MB CPU
LeViT ONNX w8a16 Snapdragon® 8 Elite For Galaxy Mobile 0.517 ms 0 - 58 MB NPU
LeViT ONNX w8a16 Qualcomm® QCS9075 1.25 ms 0 - 45 MB NPU
LeViT ONNX w8a16 Qualcomm® QCS7790 28.139 ms 13 - 23 MB CPU
LeViT ONNX w8a16 Qualcomm® QCS8750 0.517 ms 0 - 58 MB NPU
LeViT ONNX w8a16 Qualcomm® QCS7181 1.053 ms 149 - 149 MB NPU
LeViT ONNX w8a16_mixed_int16 Snapdragon® 8 Elite Gen 5 Mobile 0.479 ms 0 - 61 MB NPU
LeViT ONNX w8a16_mixed_int16 Snapdragon® X2 Elite 0.51 ms 212 - 212 MB NPU
LeViT ONNX w8a16_mixed_int16 Snapdragon® X Elite 1.1 ms 149 - 149 MB NPU
LeViT ONNX w8a16_mixed_int16 Snapdragon® 8 Gen 3 Mobile 0.716 ms 0 - 91 MB NPU
LeViT ONNX w8a16_mixed_int16 Qualcomm® QCS6490 58.283 ms 13 - 18 MB CPU
LeViT ONNX w8a16_mixed_int16 Qualcomm® QCS8550 (Proxy) 1.06 ms 0 - 46 MB NPU
LeViT ONNX w8a16_mixed_int16 Qualcomm® QCM6690 31.913 ms 12 - 21 MB CPU
LeViT ONNX w8a16_mixed_int16 Snapdragon® 8 Elite For Galaxy Mobile 0.543 ms 0 - 59 MB NPU
LeViT ONNX w8a16_mixed_int16 Qualcomm® QCS9075 1.33 ms 0 - 45 MB NPU
LeViT ONNX w8a16_mixed_int16 Qualcomm® QCS8750 0.543 ms 0 - 59 MB NPU
LeViT ONNX w8a16_mixed_int16 Qualcomm® QCS7181 1.1 ms 149 - 149 MB NPU
LeViT QNN_DLC float Snapdragon® 8 Elite Gen 5 Mobile 0.719 ms 1 - 61 MB NPU
LeViT QNN_DLC float Snapdragon® X2 Elite 0.94 ms 1 - 1 MB NPU
LeViT QNN_DLC float Snapdragon® X Elite 1.812 ms 1 - 1 MB NPU
LeViT QNN_DLC float Snapdragon® 8 Gen 3 Mobile 1.066 ms 0 - 84 MB NPU
LeViT QNN_DLC float Qualcomm® QCS8275 3.853 ms 0 - 57 MB NPU
LeViT QNN_DLC float Qualcomm® QCS8550 (Proxy) 1.61 ms 1 - 34 MB NPU
LeViT QNN_DLC float Qualcomm® SA8775P 1.991 ms 1 - 60 MB NPU
LeViT QNN_DLC float Qualcomm® SA8650P 1.991 ms 1 - 60 MB NPU
LeViT QNN_DLC float Qualcomm® SA8255P 1.991 ms 1 - 60 MB NPU
LeViT QNN_DLC float Qualcomm® QCS8450 (Proxy) 2.379 ms 0 - 83 MB NPU
LeViT QNN_DLC float Qualcomm® SA7255P 3.853 ms 0 - 57 MB NPU
LeViT QNN_DLC float Snapdragon® 8 Elite For Galaxy Mobile 0.838 ms 0 - 59 MB NPU
LeViT QNN_DLC float Qualcomm® QCS9075 1.906 ms 1 - 3 MB NPU
LeViT QNN_DLC float Qualcomm® SA8295P 2.375 ms 0 - 52 MB NPU
LeViT QNN_DLC float Qualcomm® QCS8750 0.838 ms 0 - 59 MB NPU
LeViT QNN_DLC float Qualcomm® QCS7181 1.812 ms 1 - 1 MB NPU
LeViT QNN_DLC w8a16 Snapdragon® 8 Elite Gen 5 Mobile 0.564 ms 0 - 54 MB NPU
LeViT QNN_DLC w8a16 Snapdragon® X2 Elite 0.792 ms 0 - 0 MB NPU
LeViT QNN_DLC w8a16 Snapdragon® X Elite 1.584 ms 0 - 0 MB NPU
LeViT QNN_DLC w8a16 Snapdragon® 8 Gen 3 Mobile 0.934 ms 0 - 72 MB NPU
LeViT QNN_DLC w8a16 Qualcomm® QCS8275 2.756 ms 0 - 50 MB NPU
LeViT QNN_DLC w8a16 Qualcomm® QCS8550 (Proxy) 1.39 ms 0 - 2 MB NPU
LeViT QNN_DLC w8a16 Qualcomm® SA8775P 1.682 ms 0 - 52 MB NPU
LeViT QNN_DLC w8a16 Qualcomm® SA8650P 1.682 ms 0 - 52 MB NPU
LeViT QNN_DLC w8a16 Qualcomm® SA8255P 1.682 ms 0 - 52 MB NPU
LeViT QNN_DLC w8a16 Qualcomm® QCM6690 5.526 ms 0 - 173 MB NPU
LeViT QNN_DLC w8a16 Qualcomm® SA7255P 2.756 ms 0 - 50 MB NPU
LeViT QNN_DLC w8a16 Snapdragon® 7 Gen 4 Mobile 1.416 ms 0 - 50 MB NPU
LeViT QNN_DLC w8a16 Snapdragon® 8 Elite For Galaxy Mobile 0.659 ms 0 - 47 MB NPU
LeViT QNN_DLC w8a16 Qualcomm® QCS9075 1.717 ms 0 - 2 MB NPU
LeViT QNN_DLC w8a16 Qualcomm® QCS7790 1.416 ms 0 - 50 MB NPU
LeViT QNN_DLC w8a16 Qualcomm® QCS8750 0.659 ms 0 - 47 MB NPU
LeViT QNN_DLC w8a16 Qualcomm® QCS7181 1.584 ms 0 - 0 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Snapdragon® 8 Elite Gen 5 Mobile 0.606 ms 0 - 55 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Snapdragon® X2 Elite 0.817 ms 0 - 0 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Snapdragon® X Elite 1.67 ms 0 - 0 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Snapdragon® 8 Gen 3 Mobile 0.968 ms 0 - 76 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Qualcomm® QCS8275 2.892 ms 0 - 51 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Qualcomm® QCS8550 (Proxy) 1.431 ms 0 - 2 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Qualcomm® SA8775P 1.748 ms 0 - 52 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Qualcomm® SA8650P 1.748 ms 0 - 52 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Qualcomm® SA8255P 1.748 ms 0 - 52 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Qualcomm® QCM6690 5.953 ms 0 - 176 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Qualcomm® SA7255P 2.892 ms 0 - 51 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Snapdragon® 7 Gen 4 Mobile 1.434 ms 0 - 52 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Snapdragon® 8 Elite For Galaxy Mobile 0.699 ms 0 - 49 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Qualcomm® QCS9075 1.71 ms 2 - 4 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Qualcomm® QCS7790 1.434 ms 0 - 52 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Qualcomm® QCS8750 0.699 ms 0 - 49 MB NPU
LeViT QNN_DLC w8a16_mixed_int16 Qualcomm® QCS7181 1.67 ms 0 - 0 MB NPU
LeViT TFLITE float Snapdragon® 8 Elite Gen 5 Mobile 0.658 ms 0 - 68 MB NPU
LeViT TFLITE float Snapdragon® 8 Gen 3 Mobile 1.044 ms 0 - 88 MB NPU
LeViT TFLITE float Qualcomm® QCS8275 3.997 ms 0 - 62 MB NPU
LeViT TFLITE float Qualcomm® QCS8550 (Proxy) 1.527 ms 0 - 4 MB NPU
LeViT TFLITE float Qualcomm® SA8775P 1.984 ms 0 - 62 MB NPU
LeViT TFLITE float Qualcomm® SA8650P 1.984 ms 0 - 62 MB NPU
LeViT TFLITE float Qualcomm® SA8255P 1.984 ms 0 - 62 MB NPU
LeViT TFLITE float Qualcomm® QCS8450 (Proxy) 2.362 ms 0 - 88 MB NPU
LeViT TFLITE float Qualcomm® SA7255P 3.997 ms 0 - 62 MB NPU
LeViT TFLITE float Snapdragon® 8 Elite For Galaxy Mobile 0.78 ms 0 - 70 MB NPU
LeViT TFLITE float Qualcomm® QCS9075 1.857 ms 0 - 19 MB NPU
LeViT TFLITE float Qualcomm® SA8295P 2.34 ms 0 - 53 MB NPU
LeViT TFLITE float Qualcomm® QCS8750 0.78 ms 0 - 70 MB NPU

License

  • The license for the original implementation of LeViT 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/LeViT