Hugging Face's logo

qualcomm
/
BiseNet

Image Segmentation
PyTorch
LiteRT
ONNX
real_time
android
Model card Files Community

BiseNet: Optimized for Mobile Deployment

Segment images or video by class in real-time on device

BiSeNet (Bilateral Segmentation Network) is a novel architecture designed for real-time semantic segmentation. It addresses the challenge of balancing spatial resolution and receptive field by employing a Spatial Path to preserve high-resolution features and a context path to capture sufficient receptive field.

This model is an implementation of BiseNet found here.

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

Model Details

  • Model Type: Semantic segmentation
  • Model Stats:
    • Model checkpoint: best_dice_loss_miou_0.655.pth
    • Inference latency: RealTime
    • Input resolution: 720x960
    • Number of parameters: 12.0M
    • Model size: 45.7 MB
Model Device Chipset Target Runtime Inference Time (ms) Peak Memory Range (MB) Precision Primary Compute Unit Target Model
BiseNet Samsung Galaxy S23 Snapdragon® 8 Gen 2 TFLITE 28.162 ms 12 - 46 MB FP16 NPU BiseNet.tflite
BiseNet Samsung Galaxy S23 Snapdragon® 8 Gen 2 QNN 26.93 ms 8 - 11 MB FP16 NPU BiseNet.so
BiseNet Samsung Galaxy S23 Snapdragon® 8 Gen 2 ONNX 37.632 ms 63 - 124 MB FP16 NPU BiseNet.onnx
BiseNet Samsung Galaxy S24 Snapdragon® 8 Gen 3 TFLITE 21.228 ms 31 - 78 MB FP16 NPU BiseNet.tflite
BiseNet Samsung Galaxy S24 Snapdragon® 8 Gen 3 QNN 20.423 ms 8 - 26 MB FP16 NPU BiseNet.so
BiseNet Samsung Galaxy S24 Snapdragon® 8 Gen 3 ONNX 27.205 ms 65 - 98 MB FP16 NPU BiseNet.onnx
BiseNet Snapdragon 8 Elite QRD Snapdragon® 8 Elite TFLITE 19.018 ms 30 - 58 MB FP16 NPU BiseNet.tflite
BiseNet Snapdragon 8 Elite QRD Snapdragon® 8 Elite QNN 19.088 ms 8 - 34 MB FP16 NPU Use Export Script
BiseNet Snapdragon 8 Elite QRD Snapdragon® 8 Elite ONNX 25.844 ms 73 - 103 MB FP16 NPU BiseNet.onnx
BiseNet SA7255P ADP SA7255P TFLITE 485.254 ms 30 - 56 MB FP16 NPU BiseNet.tflite
BiseNet SA7255P ADP SA7255P QNN 484.503 ms 1 - 11 MB FP16 NPU Use Export Script
BiseNet SA8255 (Proxy) SA8255P Proxy TFLITE 28.306 ms 32 - 109 MB FP16 NPU BiseNet.tflite
BiseNet SA8255 (Proxy) SA8255P Proxy QNN 26.505 ms 8 - 10 MB FP16 NPU Use Export Script
BiseNet SA8295P ADP SA8295P TFLITE 37.834 ms 32 - 58 MB FP16 NPU BiseNet.tflite
BiseNet SA8295P ADP SA8295P QNN 35.674 ms 3 - 20 MB FP16 NPU Use Export Script
BiseNet SA8650 (Proxy) SA8650P Proxy TFLITE 28.145 ms 7 - 27 MB FP16 NPU BiseNet.tflite
BiseNet SA8650 (Proxy) SA8650P Proxy QNN 27.056 ms 8 - 10 MB FP16 NPU Use Export Script
BiseNet SA8775P ADP SA8775P TFLITE 41.242 ms 32 - 58 MB FP16 NPU BiseNet.tflite
BiseNet SA8775P ADP SA8775P QNN 39.463 ms 2 - 12 MB FP16 NPU Use Export Script
BiseNet QCS8275 (Proxy) QCS8275 Proxy TFLITE 485.254 ms 30 - 56 MB FP16 NPU BiseNet.tflite
BiseNet QCS8275 (Proxy) QCS8275 Proxy QNN 484.503 ms 1 - 11 MB FP16 NPU Use Export Script
BiseNet QCS8550 (Proxy) QCS8550 Proxy TFLITE 27.951 ms 13 - 72 MB FP16 NPU BiseNet.tflite
BiseNet QCS8550 (Proxy) QCS8550 Proxy QNN 26.856 ms 8 - 10 MB FP16 NPU Use Export Script
BiseNet QCS9075 (Proxy) QCS9075 Proxy TFLITE 41.242 ms 32 - 58 MB FP16 NPU BiseNet.tflite
BiseNet QCS9075 (Proxy) QCS9075 Proxy QNN 39.463 ms 2 - 12 MB FP16 NPU Use Export Script
BiseNet QCS8450 (Proxy) QCS8450 Proxy TFLITE 35.572 ms 32 - 80 MB FP16 NPU BiseNet.tflite
BiseNet QCS8450 (Proxy) QCS8450 Proxy QNN 35.517 ms 8 - 40 MB FP16 NPU Use Export Script
BiseNet Snapdragon X Elite CRD Snapdragon® X Elite QNN 25.281 ms 8 - 8 MB FP16 NPU Use Export Script
BiseNet Snapdragon X Elite CRD Snapdragon® X Elite ONNX 36.024 ms 66 - 66 MB FP16 NPU BiseNet.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.bisenet.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.bisenet.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.bisenet.export

Profiling Results
------------------------------------------------------------
BiseNet
Device                          : Samsung Galaxy S23 (13)
Runtime                         : TFLITE                 
Estimated inference time (ms)   : 28.2                   
Estimated peak memory usage (MB): [12, 46]               
Total # Ops                     : 63                     
Compute Unit(s)                 : NPU (63 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.bisenet 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.bisenet.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.bisenet.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 BiseNet's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

  • The license for the original implementation of BiseNet can be found [here](This model's original implementation does not provide a LICENSE.).
  • 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 Segmentation
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-segmentation models for pytorch library.