ResNet101: Optimized for Mobile Deployment

Imagenet classifier and general purpose backbone

ResNet101 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 ResNet101 found here.

This repository provides scripts to run ResNet101 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: 44.5M
    • Model size (float): 170 MB
    • Model size (w8a8): 43.9 MB
Model Precision Device Chipset Target Runtime Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit Target Model
ResNet101 float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) TFLITE 18.303 ms 0 - 107 MB NPU ResNet101.tflite
ResNet101 float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) QNN_DLC 18.249 ms 1 - 39 MB NPU ResNet101.dlc
ResNet101 float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) TFLITE 4.734 ms 0 - 110 MB NPU ResNet101.tflite
ResNet101 float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) QNN_DLC 5.479 ms 1 - 32 MB NPU ResNet101.dlc
ResNet101 float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) TFLITE 3.349 ms 0 - 412 MB NPU ResNet101.tflite
ResNet101 float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) QNN_DLC 3.228 ms 1 - 17 MB NPU ResNet101.dlc
ResNet101 float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) TFLITE 5.511 ms 0 - 108 MB NPU ResNet101.tflite
ResNet101 float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) QNN_DLC 5.337 ms 0 - 39 MB NPU ResNet101.dlc
ResNet101 float SA7255P ADP Qualcomm® SA7255P TFLITE 18.303 ms 0 - 107 MB NPU ResNet101.tflite
ResNet101 float SA7255P ADP Qualcomm® SA7255P QNN_DLC 18.249 ms 1 - 39 MB NPU ResNet101.dlc
ResNet101 float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) TFLITE 3.361 ms 0 - 465 MB NPU ResNet101.tflite
ResNet101 float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) QNN_DLC 3.254 ms 0 - 22 MB NPU ResNet101.dlc
ResNet101 float SA8295P ADP Qualcomm® SA8295P TFLITE 5.583 ms 0 - 81 MB NPU ResNet101.tflite
ResNet101 float SA8295P ADP Qualcomm® SA8295P QNN_DLC 5.489 ms 1 - 29 MB NPU ResNet101.dlc
ResNet101 float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) TFLITE 3.332 ms 0 - 291 MB NPU ResNet101.tflite
ResNet101 float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) QNN_DLC 3.244 ms 0 - 17 MB NPU ResNet101.dlc
ResNet101 float SA8775P ADP Qualcomm® SA8775P TFLITE 5.511 ms 0 - 108 MB NPU ResNet101.tflite
ResNet101 float SA8775P ADP Qualcomm® SA8775P QNN_DLC 5.337 ms 0 - 39 MB NPU ResNet101.dlc
ResNet101 float Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile TFLITE 3.326 ms 0 - 31 MB NPU ResNet101.tflite
ResNet101 float Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile QNN_DLC 3.247 ms 0 - 17 MB NPU ResNet101.dlc
ResNet101 float Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile ONNX 3.295 ms 0 - 205 MB NPU ResNet101.onnx
ResNet101 float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile TFLITE 2.407 ms 0 - 136 MB NPU ResNet101.tflite
ResNet101 float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile QNN_DLC 2.386 ms 0 - 46 MB NPU ResNet101.dlc
ResNet101 float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile ONNX 2.405 ms 0 - 49 MB NPU ResNet101.onnx
ResNet101 float Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile TFLITE 2.292 ms 0 - 112 MB NPU ResNet101.tflite
ResNet101 float Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile QNN_DLC 2.241 ms 0 - 44 MB NPU ResNet101.dlc
ResNet101 float Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile ONNX 2.39 ms 1 - 42 MB NPU ResNet101.onnx
ResNet101 float Snapdragon X Elite CRD Snapdragon® X Elite QNN_DLC 3.827 ms 231 - 231 MB NPU ResNet101.dlc
ResNet101 float Snapdragon X Elite CRD Snapdragon® X Elite ONNX 3.361 ms 87 - 87 MB NPU ResNet101.onnx
ResNet101 w8a8 QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) TFLITE 2.708 ms 0 - 43 MB NPU ResNet101.tflite
ResNet101 w8a8 QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) QNN_DLC 2.931 ms 0 - 46 MB NPU ResNet101.dlc
ResNet101 w8a8 QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) TFLITE 1.39 ms 0 - 94 MB NPU ResNet101.tflite
ResNet101 w8a8 QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) QNN_DLC 1.806 ms 0 - 83 MB NPU ResNet101.dlc
ResNet101 w8a8 QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) TFLITE 1.158 ms 0 - 235 MB NPU ResNet101.tflite
ResNet101 w8a8 QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) QNN_DLC 1.274 ms 0 - 232 MB NPU ResNet101.dlc
ResNet101 w8a8 QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) TFLITE 1.331 ms 0 - 43 MB NPU ResNet101.tflite
ResNet101 w8a8 QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) QNN_DLC 1.431 ms 0 - 46 MB NPU ResNet101.dlc
ResNet101 w8a8 RB3 Gen 2 (Proxy) Qualcomm® QCS6490 (Proxy) TFLITE 4.359 ms 0 - 76 MB NPU ResNet101.tflite
ResNet101 w8a8 RB3 Gen 2 (Proxy) Qualcomm® QCS6490 (Proxy) QNN_DLC 6.309 ms 0 - 78 MB NPU ResNet101.dlc
ResNet101 w8a8 RB5 (Proxy) Qualcomm® QCS8250 (Proxy) TFLITE 17.096 ms 0 - 3 MB NPU ResNet101.tflite
ResNet101 w8a8 SA7255P ADP Qualcomm® SA7255P TFLITE 2.708 ms 0 - 43 MB NPU ResNet101.tflite
ResNet101 w8a8 SA7255P ADP Qualcomm® SA7255P QNN_DLC 2.931 ms 0 - 46 MB NPU ResNet101.dlc
ResNet101 w8a8 SA8255 (Proxy) Qualcomm® SA8255P (Proxy) TFLITE 1.161 ms 0 - 233 MB NPU ResNet101.tflite
ResNet101 w8a8 SA8255 (Proxy) Qualcomm® SA8255P (Proxy) QNN_DLC 1.274 ms 0 - 206 MB NPU ResNet101.dlc
ResNet101 w8a8 SA8295P ADP Qualcomm® SA8295P TFLITE 1.782 ms 0 - 46 MB NPU ResNet101.tflite
ResNet101 w8a8 SA8295P ADP Qualcomm® SA8295P QNN_DLC 1.881 ms 0 - 49 MB NPU ResNet101.dlc
ResNet101 w8a8 SA8650 (Proxy) Qualcomm® SA8650P (Proxy) TFLITE 1.162 ms 0 - 235 MB NPU ResNet101.tflite
ResNet101 w8a8 SA8650 (Proxy) Qualcomm® SA8650P (Proxy) QNN_DLC 1.274 ms 0 - 228 MB NPU ResNet101.dlc
ResNet101 w8a8 SA8775P ADP Qualcomm® SA8775P TFLITE 1.331 ms 0 - 43 MB NPU ResNet101.tflite
ResNet101 w8a8 SA8775P ADP Qualcomm® SA8775P QNN_DLC 1.431 ms 0 - 46 MB NPU ResNet101.dlc
ResNet101 w8a8 Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile TFLITE 1.173 ms 0 - 235 MB NPU ResNet101.tflite
ResNet101 w8a8 Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile QNN_DLC 1.274 ms 0 - 230 MB NPU ResNet101.dlc
ResNet101 w8a8 Samsung Galaxy S23 Snapdragon® 8 Gen 2 Mobile ONNX 24.537 ms 3 - 150 MB NPU ResNet101.onnx
ResNet101 w8a8 Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile TFLITE 0.883 ms 0 - 90 MB NPU ResNet101.tflite
ResNet101 w8a8 Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile QNN_DLC 0.975 ms 0 - 79 MB NPU ResNet101.dlc
ResNet101 w8a8 Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile ONNX 20.199 ms 11 - 1672 MB NPU ResNet101.onnx
ResNet101 w8a8 Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile TFLITE 0.795 ms 0 - 56 MB NPU ResNet101.tflite
ResNet101 w8a8 Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile QNN_DLC 0.858 ms 0 - 52 MB NPU ResNet101.dlc
ResNet101 w8a8 Snapdragon 8 Elite QRD Snapdragon® 8 Elite Mobile ONNX 17.414 ms 12 - 691 MB NPU ResNet101.onnx
ResNet101 w8a8 Snapdragon X Elite CRD Snapdragon® X Elite QNN_DLC 1.274 ms 204 - 204 MB NPU ResNet101.dlc
ResNet101 w8a8 Snapdragon X Elite CRD Snapdragon® X Elite ONNX 24.279 ms 51 - 51 MB NPU ResNet101.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.resnet101.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.resnet101.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.resnet101.export
Profiling Results
------------------------------------------------------------
ResNet101
Device                          : cs_8275 (ANDROID 14)                 
Runtime                         : TFLITE                               
Estimated inference time (ms)   : 18.3                                 
Estimated peak memory usage (MB): [0, 107]                             
Total # Ops                     : 147                                  
Compute Unit(s)                 : npu (147 ops) gpu (0 ops) cpu (0 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.resnet101 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.resnet101.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.resnet101.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 ResNet101's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

  • The license for the original implementation of ResNet101 can be found here.
  • The license for the compiled assets for on-device deployment can be found here

References

Community

Downloads last month
135
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support