Hugging Face's logo

qualcomm
/
NASNet

Image Classification
PyTorch
bu_auto
android
Model card Files
xet
 Community
NASNet / README.md
qaihm-bot's picture
qaihm-bot
v0.58.0
e6d6827 verified
preview code
|
Raw
History Blame Contribute Delete
8.58 kB
metadata 
library_name: pytorch
license: other
tags:
  - bu_auto
  - android
pipeline_tag: image-classification

NASNet: Optimized for Qualcomm Devices

NASNet is a CNN-based architecture discovered via Neural Architecture Search (NAS) that can classify images from the Imagenet dataset.

This is based on the implementation of NASNet 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 1.25.0 Download
ONNX w8a8_mixed_fp16 Universal QAIRT 2.45, ONNX Runtime 1.25.0 Download
QNN_DLC float Universal QAIRT 2.45 Download
TFLITE float Universal QAIRT 2.45 Download

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

Model Details

Model Type: Model_use_case.image_classification

Model Stats:

  • Model checkpoint: nasnetalarge.tf_in1k
  • Input resolution: 224x224
  • GMACs: 5.9
  • Activations (M): 19.4
  • Number of parameters: 88.7M
  • Model size (float): 338 MB

Performance Summary

Model Runtime Precision Chipset Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit
NASNet ONNX float Snapdragon® X2 Elite 9.577 ms 1 - 1 MB NPU
NASNet ONNX float Snapdragon® X Elite 18.942 ms 188 - 188 MB NPU
NASNet ONNX float Snapdragon® 8 Gen 3 Mobile 14.232 ms 1 - 571 MB NPU
NASNet ONNX float Snapdragon® 8 Gen 1 Mobile 52.77 ms 2 - 518 MB NPU
NASNet ONNX float Qualcomm® Dragonwing™ QCS8550 (Proxy) 18.818 ms 0 - 194 MB NPU
NASNet ONNX float Qualcomm® QCS8450 52.77 ms 2 - 518 MB NPU
NASNet ONNX float Qualcomm® Dragonwing™ IQ-9075 27.975 ms 1 - 5 MB NPU
NASNet ONNX float Snapdragon® 8 Elite Gen 5 Mobile 9.504 ms 1 - 416 MB NPU
NASNet ONNX float Snapdragon® 8 Elite Mobile 11.844 ms 1 - 405 MB NPU
NASNet ONNX float Qualcomm® Dragonwing™ Q-8750 11.844 ms 1 - 405 MB NPU
NASNet ONNX float Qualcomm® Dragonwing™ IQ-X7181 18.942 ms 188 - 188 MB NPU
NASNet ONNX w8a8_mixed_fp16 Snapdragon® X2 Elite 4.672 ms 1 - 1 MB NPU
NASNet ONNX w8a8_mixed_fp16 Snapdragon® X Elite 9.409 ms 99 - 99 MB NPU
NASNet ONNX w8a8_mixed_fp16 Snapdragon® 8 Gen 3 Mobile 6.52 ms 0 - 618 MB NPU
NASNet ONNX w8a8_mixed_fp16 Snapdragon® 8 Gen 1 Mobile 19.308 ms 0 - 598 MB NPU
NASNet ONNX w8a8_mixed_fp16 Qualcomm® Dragonwing™ QCS8550 (Proxy) 9.23 ms 1 - 4 MB NPU
NASNet ONNX w8a8_mixed_fp16 Qualcomm® QCS8450 19.308 ms 0 - 598 MB NPU
NASNet ONNX w8a8_mixed_fp16 Qualcomm® Dragonwing™ IQ-9075 9.466 ms 0 - 3 MB NPU
NASNet ONNX w8a8_mixed_fp16 Snapdragon® 8 Elite Gen 5 Mobile 4.717 ms 0 - 482 MB NPU
NASNet ONNX w8a8_mixed_fp16 Snapdragon® 8 Elite Mobile 5.338 ms 0 - 488 MB NPU
NASNet ONNX w8a8_mixed_fp16 Qualcomm® Dragonwing™ Q-8750 5.338 ms 0 - 488 MB NPU
NASNet ONNX w8a8_mixed_fp16 Qualcomm® Dragonwing™ IQ-X7181 9.409 ms 99 - 99 MB NPU
NASNet QNN_DLC float Snapdragon® X2 Elite 8.9 ms 1 - 1 MB NPU
NASNet QNN_DLC float Snapdragon® X Elite 16.762 ms 1 - 1 MB NPU
NASNet QNN_DLC float Snapdragon® 8 Gen 3 Mobile 12.288 ms 1 - 516 MB NPU
NASNet QNN_DLC float Snapdragon® 8 Gen 1 Mobile 48.631 ms 0 - 474 MB NPU
NASNet QNN_DLC float Qualcomm® QCS8275 89.179 ms 2 - 351 MB NPU
NASNet QNN_DLC float Qualcomm® Dragonwing™ QCS8550 (Proxy) 16.416 ms 1 - 3 MB NPU
NASNet QNN_DLC float Qualcomm® SA8775P 24.423 ms 1 - 354 MB NPU
NASNet QNN_DLC float Qualcomm® SA8650P 24.423 ms 1 - 354 MB NPU
NASNet QNN_DLC float Qualcomm® SA8255P 24.423 ms 1 - 354 MB NPU
NASNet QNN_DLC float Qualcomm® QCS8450 48.631 ms 0 - 474 MB NPU
NASNet QNN_DLC float Qualcomm® Dragonwing™ IQ-9075 24.184 ms 1 - 4 MB NPU
NASNet QNN_DLC float Snapdragon® 8 Elite Gen 5 Mobile 7.823 ms 0 - 367 MB NPU
NASNet QNN_DLC float Qualcomm® SA7255P 89.179 ms 2 - 351 MB NPU
NASNet QNN_DLC float Snapdragon® 8 Elite Mobile 9.94 ms 1 - 352 MB NPU
NASNet QNN_DLC float Qualcomm® SA8295P 38.611 ms 1 - 310 MB NPU
NASNet QNN_DLC float Qualcomm® Dragonwing™ Q-8750 9.94 ms 1 - 352 MB NPU
NASNet QNN_DLC float Qualcomm® Dragonwing™ IQ-X7181 16.762 ms 1 - 1 MB NPU
NASNet TFLITE float Snapdragon® 8 Gen 3 Mobile 12.568 ms 0 - 691 MB NPU
NASNet TFLITE float Snapdragon® 8 Gen 1 Mobile 50.677 ms 0 - 653 MB NPU
NASNet TFLITE float Qualcomm® QCS8275 89.649 ms 0 - 538 MB NPU
NASNet TFLITE float Qualcomm® Dragonwing™ QCS8550 (Proxy) 16.559 ms 7 - 10 MB NPU
NASNet TFLITE float Qualcomm® SA8775P 24.741 ms 0 - 538 MB NPU
NASNet TFLITE float Qualcomm® SA8650P 24.741 ms 0 - 538 MB NPU
NASNet TFLITE float Qualcomm® SA8255P 24.741 ms 0 - 538 MB NPU
NASNet TFLITE float Qualcomm® QCS8450 50.677 ms 0 - 653 MB NPU
NASNet TFLITE float Qualcomm® Dragonwing™ IQ-9075 25.073 ms 0 - 192 MB NPU
NASNet TFLITE float Snapdragon® 8 Elite Gen 5 Mobile 8.24 ms 0 - 550 MB NPU
NASNet TFLITE float Qualcomm® SA7255P 89.649 ms 0 - 538 MB NPU
NASNet TFLITE float Snapdragon® 8 Elite Mobile 10.325 ms 0 - 532 MB NPU
NASNet TFLITE float Qualcomm® SA8295P 40.074 ms 0 - 486 MB NPU
NASNet TFLITE float Qualcomm® Dragonwing™ Q-8750 10.325 ms 0 - 532 MB NPU

License

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

References

Community