Hugging Face's logo

qualcomm
/
ConvNext-Tiny

Image Classification
PyTorch
TF Lite
ONNX
android
Model card Files Community

ConvNext-Tiny: Optimized for Mobile Deployment

Imagenet classifier and general purpose backbone

ConvNextTiny 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 model is an implementation of ConvNext-Tiny found here.

This repository provides scripts to run ConvNext-Tiny on Qualcomm® devices. More details on model performance across various devices, can be found here.

Model Details

  • Model Type: Image classification
  • Model Stats:
    • Model checkpoint: Imagenet
    • Input resolution: 224x224
    • Number of parameters: 28.6M
    • Model size: 109 MB
Model Device Chipset Target Runtime Inference Time (ms) Peak Memory Range (MB) Precision Primary Compute Unit Target Model
ConvNext-Tiny Samsung Galaxy S23 Snapdragon® 8 Gen 2 TFLITE 3.328 ms 0 - 289 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny Samsung Galaxy S23 Snapdragon® 8 Gen 2 QNN 3.676 ms 0 - 3 MB FP16 NPU ConvNext-Tiny.so
ConvNext-Tiny Samsung Galaxy S23 Snapdragon® 8 Gen 2 ONNX 13.475 ms 0 - 149 MB FP16 NPU ConvNext-Tiny.onnx
ConvNext-Tiny Samsung Galaxy S24 Snapdragon® 8 Gen 3 TFLITE 2.452 ms 0 - 37 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny Samsung Galaxy S24 Snapdragon® 8 Gen 3 QNN 2.636 ms 1 - 19 MB FP16 NPU ConvNext-Tiny.so
ConvNext-Tiny Samsung Galaxy S24 Snapdragon® 8 Gen 3 ONNX 9.64 ms 1 - 61 MB FP16 NPU ConvNext-Tiny.onnx
ConvNext-Tiny Snapdragon 8 Elite QRD Snapdragon® 8 Elite TFLITE 2.115 ms 0 - 41 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny Snapdragon 8 Elite QRD Snapdragon® 8 Elite QNN 2.438 ms 1 - 40 MB FP16 NPU Use Export Script
ConvNext-Tiny Snapdragon 8 Elite QRD Snapdragon® 8 Elite ONNX 7.405 ms 1 - 67 MB FP16 NPU ConvNext-Tiny.onnx
ConvNext-Tiny SA7255P ADP SA7255P TFLITE 96.654 ms 0 - 35 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny SA7255P ADP SA7255P QNN 97.217 ms 1 - 8 MB FP16 NPU Use Export Script
ConvNext-Tiny SA8255 (Proxy) SA8255P Proxy TFLITE 3.338 ms 0 - 259 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny SA8255 (Proxy) SA8255P Proxy QNN 3.663 ms 1 - 3 MB FP16 NPU Use Export Script
ConvNext-Tiny SA8295P ADP SA8295P TFLITE 11.113 ms 0 - 34 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny SA8295P ADP SA8295P QNN 9.669 ms 1 - 11 MB FP16 NPU Use Export Script
ConvNext-Tiny SA8650 (Proxy) SA8650P Proxy TFLITE 3.341 ms 0 - 299 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny SA8650 (Proxy) SA8650P Proxy QNN 3.676 ms 1 - 3 MB FP16 NPU Use Export Script
ConvNext-Tiny SA8775P ADP SA8775P TFLITE 5.724 ms 0 - 36 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny SA8775P ADP SA8775P QNN 6.172 ms 1 - 8 MB FP16 NPU Use Export Script
ConvNext-Tiny QCS8275 (Proxy) QCS8275 Proxy TFLITE 96.654 ms 0 - 35 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny QCS8275 (Proxy) QCS8275 Proxy QNN 97.217 ms 1 - 8 MB FP16 NPU Use Export Script
ConvNext-Tiny QCS8550 (Proxy) QCS8550 Proxy TFLITE 3.328 ms 0 - 309 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny QCS8550 (Proxy) QCS8550 Proxy QNN 3.639 ms 1 - 3 MB FP16 NPU Use Export Script
ConvNext-Tiny QCS9075 (Proxy) QCS9075 Proxy TFLITE 5.724 ms 0 - 36 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny QCS9075 (Proxy) QCS9075 Proxy QNN 6.172 ms 1 - 8 MB FP16 NPU Use Export Script
ConvNext-Tiny QCS8450 (Proxy) QCS8450 Proxy TFLITE 9.877 ms 0 - 36 MB FP16 NPU ConvNext-Tiny.tflite
ConvNext-Tiny QCS8450 (Proxy) QCS8450 Proxy QNN 10.364 ms 1 - 36 MB FP16 NPU Use Export Script
ConvNext-Tiny Snapdragon X Elite CRD Snapdragon® X Elite QNN 3.934 ms 1 - 1 MB FP16 NPU Use Export Script
ConvNext-Tiny Snapdragon X Elite CRD Snapdragon® X Elite ONNX 16.203 ms 59 - 59 MB FP16 NPU ConvNext-Tiny.onnx

Installation

Install the package via pip:

pip install qai-hub-models

Configure Qualcomm® AI Hub to run this model on a cloud-hosted device

Sign-in to Qualcomm® AI Hub with your Qualcomm® ID. Once signed in navigate to Account -> Settings -> API Token.

With this API token, you can configure your client to run models on the cloud hosted devices.

qai-hub configure --api_token API_TOKEN

Navigate to docs for more information.

Demo off target

The package contains a simple end-to-end demo that downloads pre-trained weights and runs this model on a sample input.

python -m qai_hub_models.models.convnext_tiny.demo

The above demo runs a reference implementation of pre-processing, model inference, and post processing.

NOTE: If you want running in a Jupyter Notebook or Google Colab like environment, please add the following to your cell (instead of the above).

%run -m qai_hub_models.models.convnext_tiny.demo

Run model on a cloud-hosted device

In addition to the demo, you can also run the model on a cloud-hosted Qualcomm® device. This script does the following:

  • Performance check on-device on a cloud-hosted device
  • Downloads compiled assets that can be deployed on-device for Android.
  • Accuracy check between PyTorch and on-device outputs.
python -m qai_hub_models.models.convnext_tiny.export

Profiling Results
------------------------------------------------------------
ConvNext-Tiny
Device                          : Samsung Galaxy S23 (13)
Runtime                         : TFLITE                 
Estimated inference time (ms)   : 3.3                    
Estimated peak memory usage (MB): [0, 289]               
Total # Ops                     : 328                    
Compute Unit(s)                 : NPU (328 ops)

How does this work?

This export script leverages Qualcomm® AI Hub to optimize, validate, and deploy this model on-device. Lets go through each step below in detail:

Step 1: Compile model for on-device deployment

To compile a PyTorch model for on-device deployment, we first trace the model in memory using the jit.trace and then call the submit_compile_job API.

import torch

import qai_hub as hub
from qai_hub_models.models.convnext_tiny import Model

# Load the model
torch_model = Model.from_pretrained()

# Device
device = hub.Device("Samsung Galaxy S24")

# Trace model
input_shape = torch_model.get_input_spec()
sample_inputs = torch_model.sample_inputs()

pt_model = torch.jit.trace(torch_model, [torch.tensor(data[0]) for _, data in sample_inputs.items()])

# Compile model on a specific device
compile_job = hub.submit_compile_job(
    model=pt_model,
    device=device,
    input_specs=torch_model.get_input_spec(),
)

# Get target model to run on-device
target_model = compile_job.get_target_model()

Step 2: Performance profiling on cloud-hosted device

After compiling models from step 1. Models can be profiled model on-device using the target_model. Note that this scripts runs the model on a device automatically provisioned in the cloud. Once the job is submitted, you can navigate to a provided job URL to view a variety of on-device performance metrics.

profile_job = hub.submit_profile_job(
    model=target_model,
    device=device,
)

Step 3: Verify on-device accuracy

To verify the accuracy of the model on-device, you can run on-device inference on sample input data on the same cloud hosted device.

input_data = torch_model.sample_inputs()
inference_job = hub.submit_inference_job(
    model=target_model,
    device=device,
    inputs=input_data,
)
    on_device_output = inference_job.download_output_data()

With the output of the model, you can compute like PSNR, relative errors or spot check the output with expected output.

Note: This on-device profiling and inference requires access to Qualcomm® AI Hub. Sign up for access.

Run demo on a cloud-hosted device

You can also run the demo on-device.

python -m qai_hub_models.models.convnext_tiny.demo --on-device

NOTE: If you want running in a Jupyter Notebook or Google Colab like environment, please add the following to your cell (instead of the above).

%run -m qai_hub_models.models.convnext_tiny.demo -- --on-device

Deploying compiled model to Android

The models can be deployed using multiple runtimes:

  • TensorFlow Lite (.tflite export): This tutorial provides a guide to deploy the .tflite model in an Android application.

  • QNN (.so export ): This sample app provides instructions on how to use the .so shared library in an Android application.

View on Qualcomm® AI Hub

Get more details on ConvNext-Tiny's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

  • The license for the original implementation of ConvNext-Tiny can be found here.
  • The license for the compiled assets for on-device deployment 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 is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The HF Inference API does not support image-classification models for pytorch library.