Hugging Face's logo

qualcomm
/
WideResNet50

Image Classification
PyTorch
LiteRT
backbone
android
Model card Files
xet
Community

WideResNet50: Optimized for Mobile Deployment

Imagenet classifier and general purpose backbone

WideResNet50 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 WideResNet50 found here.

This repository provides scripts to run WideResNet50 on Qualcomm® devices. More details on model performance across various devices, can be found here.

Model Details

  • Model Type: Model_use_case.image_classification
  • Model Stats:
    • Model checkpoint: Imagenet
    • Input resolution: 224x224
    • Number of parameters: 68.9M
    • Model size (float): 263 MB
    • Model size (w8a8): 66.6 MB
Model Precision Device Chipset Target Runtime Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit Target Model
WideResNet50 float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) TFLITE 23.843 ms 0 - 196 MB NPU WideResNet50.tflite
WideResNet50 float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) QNN_DLC 24.34 ms 1 - 144 MB NPU WideResNet50.dlc
WideResNet50 float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) TFLITE 8.049 ms 0 - 314 MB NPU WideResNet50.tflite
WideResNet50 float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) QNN_DLC 9.041 ms 1 - 178 MB NPU WideResNet50.dlc
WideResNet50 float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) TFLITE 4.751 ms 0 - 3 MB NPU WideResNet50.tflite
WideResNet50 float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) QNN_DLC 4.784 ms 1 - 3 MB NPU WideResNet50.dlc
WideResNet50 float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) ONNX 4.808 ms 0 - 162 MB NPU WideResNet50.onnx.zip
WideResNet50 float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) TFLITE 7.139 ms 0 - 202 MB NPU WideResNet50.tflite
WideResNet50 float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) QNN_DLC 7.17 ms 1 - 145 MB NPU WideResNet50.dlc
WideResNet50 float SA7255P ADP Qualcomm® SA7255P TFLITE 23.843 ms 0 - 196 MB NPU WideResNet50.tflite
WideResNet50 float SA7255P ADP Qualcomm® SA7255P QNN_DLC 24.34 ms 1 - 144 MB NPU WideResNet50.dlc
WideResNet50 float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) TFLITE 4.749 ms 0 - 3 MB NPU WideResNet50.tflite
WideResNet50 float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) QNN_DLC 4.792 ms 1 - 3 MB NPU WideResNet50.dlc
WideResNet50 float SA8295P ADP Qualcomm® SA8295P TFLITE 7.691 ms 0 - 188 MB NPU WideResNet50.tflite
WideResNet50 float SA8295P ADP Qualcomm® SA8295P QNN_DLC 7.829 ms 1 - 133 MB NPU WideResNet50.dlc
WideResNet50 float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) TFLITE 4.765 ms 0 - 3 MB NPU WideResNet50.tflite
WideResNet50 float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) QNN_DLC 4.777 ms 1 - 3 MB NPU WideResNet50.dlc
WideResNet50 float SA8775P ADP Qualcomm® SA8775P TFLITE 7.139 ms 0 - 202 MB NPU WideResNet50.tflite
WideResNet50 float SA8775P ADP Qualcomm® SA8775P QNN_DLC 7.17 ms 1 - 145 MB NPU WideResNet50.dlc
WideResNet50 float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile TFLITE 3.636 ms 0 - 335 MB NPU WideResNet50.tflite
WideResNet50 float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile QNN_DLC 3.62 ms 0 - 197 MB NPU WideResNet50.dlc
WideResNet50 float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile ONNX 3.564 ms 0 - 167 MB NPU WideResNet50.onnx.zip
WideResNet50 float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile TFLITE 2.866 ms 0 - 199 MB NPU WideResNet50.tflite
WideResNet50 float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile QNN_DLC 2.958 ms 1 - 149 MB NPU WideResNet50.dlc
WideResNet50 float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile ONNX 3.032 ms 0 - 115 MB NPU WideResNet50.onnx.zip
WideResNet50 float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile TFLITE 2.626 ms 0 - 197 MB NPU WideResNet50.tflite
WideResNet50 float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile QNN_DLC 2.533 ms 1 - 148 MB NPU WideResNet50.dlc
WideResNet50 float Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile ONNX 2.811 ms 1 - 125 MB NPU WideResNet50.onnx.zip
WideResNet50 float Snapdragon X Elite CRD Snapdragon® X Elite QNN_DLC 4.682 ms 1 - 1 MB NPU WideResNet50.dlc
WideResNet50 float Snapdragon X Elite CRD Snapdragon® X Elite ONNX 4.482 ms 132 - 132 MB NPU WideResNet50.onnx.zip
WideResNet50 w8a8 Dragonwing Q-6690 MTP Qualcomm® Qcm6690 TFLITE 17.833 ms 0 - 190 MB NPU WideResNet50.tflite
WideResNet50 w8a8 Dragonwing Q-6690 MTP Qualcomm® Qcm6690 QNN_DLC 19.814 ms 0 - 192 MB NPU WideResNet50.dlc
WideResNet50 w8a8 Dragonwing Q-6690 MTP Qualcomm® Qcm6690 ONNX 61.363 ms 6 - 21 MB CPU WideResNet50.onnx.zip
WideResNet50 w8a8 Dragonwing RB3 Gen 2 Vision Kit Qualcomm® QCS6490 TFLITE 6.751 ms 0 - 68 MB NPU WideResNet50.tflite
WideResNet50 w8a8 Dragonwing RB3 Gen 2 Vision Kit Qualcomm® QCS6490 QNN_DLC 7.749 ms 2 - 4 MB NPU WideResNet50.dlc
WideResNet50 w8a8 Dragonwing RB3 Gen 2 Vision Kit Qualcomm® QCS6490 ONNX 77.047 ms 10 - 112 MB CPU WideResNet50.onnx.zip
WideResNet50 w8a8 QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) TFLITE 3.81 ms 0 - 143 MB NPU WideResNet50.tflite
WideResNet50 w8a8 QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) QNN_DLC 4.014 ms 0 - 144 MB NPU WideResNet50.dlc
WideResNet50 w8a8 QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) TFLITE 2.398 ms 0 - 224 MB NPU WideResNet50.tflite
WideResNet50 w8a8 QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) QNN_DLC 2.523 ms 0 - 225 MB NPU WideResNet50.dlc
WideResNet50 w8a8 QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) TFLITE 1.738 ms 0 - 2 MB NPU WideResNet50.tflite
WideResNet50 w8a8 QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) QNN_DLC 1.885 ms 0 - 2 MB NPU WideResNet50.dlc
WideResNet50 w8a8 QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) ONNX 2.052 ms 0 - 83 MB NPU WideResNet50.onnx.zip
WideResNet50 w8a8 QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) TFLITE 1.871 ms 0 - 143 MB NPU WideResNet50.tflite
WideResNet50 w8a8 QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) QNN_DLC 2.026 ms 0 - 144 MB NPU WideResNet50.dlc
WideResNet50 w8a8 RB5 (Proxy) Qualcomm® QCS8250 (Proxy) TFLITE 56.246 ms 0 - 449 MB GPU WideResNet50.tflite
WideResNet50 w8a8 RB5 (Proxy) Qualcomm® QCS8250 (Proxy) ONNX 60.446 ms 0 - 84 MB CPU WideResNet50.onnx.zip
WideResNet50 w8a8 SA7255P ADP Qualcomm® SA7255P TFLITE 3.81 ms 0 - 143 MB NPU WideResNet50.tflite
WideResNet50 w8a8 SA7255P ADP Qualcomm® SA7255P QNN_DLC 4.014 ms 0 - 144 MB NPU WideResNet50.dlc
WideResNet50 w8a8 SA8255 (Proxy) Qualcomm® SA8255P (Proxy) TFLITE 1.741 ms 0 - 2 MB NPU WideResNet50.tflite
WideResNet50 w8a8 SA8255 (Proxy) Qualcomm® SA8255P (Proxy) QNN_DLC 1.885 ms 0 - 2 MB NPU WideResNet50.dlc
WideResNet50 w8a8 SA8295P ADP Qualcomm® SA8295P TFLITE 2.616 ms 0 - 148 MB NPU WideResNet50.tflite
WideResNet50 w8a8 SA8295P ADP Qualcomm® SA8295P QNN_DLC 2.758 ms 0 - 150 MB NPU WideResNet50.dlc
WideResNet50 w8a8 SA8650 (Proxy) Qualcomm® SA8650P (Proxy) TFLITE 1.743 ms 0 - 2 MB NPU WideResNet50.tflite
WideResNet50 w8a8 SA8650 (Proxy) Qualcomm® SA8650P (Proxy) QNN_DLC 1.876 ms 0 - 2 MB NPU WideResNet50.dlc
WideResNet50 w8a8 SA8775P ADP Qualcomm® SA8775P TFLITE 1.871 ms 0 - 143 MB NPU WideResNet50.tflite
WideResNet50 w8a8 SA8775P ADP Qualcomm® SA8775P QNN_DLC 2.026 ms 0 - 144 MB NPU WideResNet50.dlc
WideResNet50 w8a8 Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile TFLITE 1.312 ms 0 - 218 MB NPU WideResNet50.tflite
WideResNet50 w8a8 Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile QNN_DLC 1.422 ms 0 - 224 MB NPU WideResNet50.dlc
WideResNet50 w8a8 Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile ONNX 1.518 ms 0 - 196 MB NPU WideResNet50.onnx.zip
WideResNet50 w8a8 Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile TFLITE 1.119 ms 0 - 148 MB NPU WideResNet50.tflite
WideResNet50 w8a8 Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile QNN_DLC 1.166 ms 0 - 145 MB NPU WideResNet50.dlc
WideResNet50 w8a8 Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile ONNX 1.32 ms 0 - 123 MB NPU WideResNet50.onnx.zip
WideResNet50 w8a8 Snapdragon 7 Gen 4 QRD Snapdragon® 7 Gen 4 Mobile TFLITE 2.78 ms 0 - 188 MB NPU WideResNet50.tflite
WideResNet50 w8a8 Snapdragon 7 Gen 4 QRD Snapdragon® 7 Gen 4 Mobile QNN_DLC 2.913 ms 0 - 193 MB NPU WideResNet50.dlc
WideResNet50 w8a8 Snapdragon 7 Gen 4 QRD Snapdragon® 7 Gen 4 Mobile ONNX 56.994 ms 8 - 23 MB CPU WideResNet50.onnx.zip
WideResNet50 w8a8 Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile TFLITE 1.049 ms 0 - 146 MB NPU WideResNet50.tflite
WideResNet50 w8a8 Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile QNN_DLC 1.082 ms 0 - 148 MB NPU WideResNet50.dlc
WideResNet50 w8a8 Snapdragon 8 Elite Gen 5 QRD Snapdragon® 8 Elite Gen5 Mobile ONNX 1.266 ms 0 - 121 MB NPU WideResNet50.onnx.zip
WideResNet50 w8a8 Snapdragon X Elite CRD Snapdragon® X Elite QNN_DLC 1.806 ms 0 - 0 MB NPU WideResNet50.dlc
WideResNet50 w8a8 Snapdragon X Elite CRD Snapdragon® X Elite ONNX 1.784 ms 67 - 67 MB NPU WideResNet50.onnx.zip

Installation

Install the package via pip:

pip install qai-hub-models

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

Sign-in to Qualcomm® AI Hub Workbench 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.wideresnet50.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.wideresnet50.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.wideresnet50.export

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.wideresnet50 import Model

# Load the model
torch_model = Model.from_pretrained()

# Device
device = hub.Device("Samsung Galaxy S25")

# 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 Workbench. 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.wideresnet50.demo --eval-mode 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.wideresnet50.demo -- --eval-mode 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 WideResNet50's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

  • The license for the original implementation of WideResNet50 can be found here.

References

Community

Downloads last month
170
Inference Providers NEW
Image Classification
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Paper for qualcomm/WideResNet50