Hugging Face's logo

qualcomm
/
BiseNet

Image Segmentation
PyTorch
LiteRT
real_time
android
Model card Files
xet
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: Model_use_case.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 (float): 45.7 MB
Model Precision Device Chipset Target Runtime Inference Time (ms) Peak Memory Range (MB) Primary Compute Unit Target Model
BiseNet float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) TFLITE 133.28 ms 32 - 67 MB NPU BiseNet.tflite
BiseNet float QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) QNN_DLC 119.385 ms 3 - 78 MB NPU BiseNet.dlc
BiseNet float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) TFLITE 55.09 ms 32 - 78 MB NPU BiseNet.tflite
BiseNet float QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) QNN_DLC 63.395 ms 8 - 60 MB NPU BiseNet.dlc
BiseNet float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) TFLITE 46.293 ms 32 - 52 MB NPU BiseNet.tflite
BiseNet float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) QNN_DLC 36.594 ms 8 - 28 MB NPU BiseNet.dlc
BiseNet float QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) ONNX 33.332 ms 71 - 92 MB NPU BiseNet.onnx.zip
BiseNet float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) TFLITE 56.28 ms 32 - 67 MB NPU BiseNet.tflite
BiseNet float QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) QNN_DLC 45.228 ms 3 - 75 MB NPU BiseNet.dlc
BiseNet float SA7255P ADP Qualcomm® SA7255P TFLITE 133.28 ms 32 - 67 MB NPU BiseNet.tflite
BiseNet float SA7255P ADP Qualcomm® SA7255P QNN_DLC 119.385 ms 3 - 78 MB NPU BiseNet.dlc
BiseNet float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) TFLITE 46.201 ms 20 - 36 MB NPU BiseNet.tflite
BiseNet float SA8255 (Proxy) Qualcomm® SA8255P (Proxy) QNN_DLC 35.511 ms 8 - 26 MB NPU BiseNet.dlc
BiseNet float SA8295P ADP Qualcomm® SA8295P TFLITE 60.739 ms 32 - 68 MB NPU BiseNet.tflite
BiseNet float SA8295P ADP Qualcomm® SA8295P QNN_DLC 51.712 ms 6 - 74 MB NPU BiseNet.dlc
BiseNet float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) TFLITE 46.212 ms 19 - 40 MB NPU BiseNet.tflite
BiseNet float SA8650 (Proxy) Qualcomm® SA8650P (Proxy) QNN_DLC 35.63 ms 8 - 29 MB NPU BiseNet.dlc
BiseNet float SA8775P ADP Qualcomm® SA8775P TFLITE 56.28 ms 32 - 67 MB NPU BiseNet.tflite
BiseNet float SA8775P ADP Qualcomm® SA8775P QNN_DLC 45.228 ms 3 - 75 MB NPU BiseNet.dlc
BiseNet float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile TFLITE 32.038 ms 31 - 80 MB NPU BiseNet.tflite
BiseNet float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile QNN_DLC 27.907 ms 8 - 67 MB NPU BiseNet.dlc
BiseNet float Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile ONNX 26.396 ms 73 - 125 MB NPU BiseNet.onnx.zip
BiseNet float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile TFLITE 26.466 ms 31 - 70 MB NPU BiseNet.tflite
BiseNet float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile QNN_DLC 20.149 ms 8 - 75 MB NPU BiseNet.dlc
BiseNet float Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile ONNX 19.317 ms 61 - 116 MB NPU BiseNet.onnx.zip
BiseNet float Snapdragon X Elite CRD Snapdragon® X Elite QNN_DLC 36.99 ms 8 - 8 MB NPU BiseNet.dlc
BiseNet float Snapdragon X Elite CRD Snapdragon® X Elite ONNX 31.619 ms 66 - 66 MB NPU BiseNet.onnx.zip
BiseNet w8a8 QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) TFLITE 24.376 ms 8 - 46 MB NPU BiseNet.tflite
BiseNet w8a8 QCS8275 (Proxy) Qualcomm® QCS8275 (Proxy) QNN_DLC 21.397 ms 2 - 51 MB NPU BiseNet.dlc
BiseNet w8a8 QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) TFLITE 16.466 ms 8 - 73 MB NPU BiseNet.tflite
BiseNet w8a8 QCS8450 (Proxy) Qualcomm® QCS8450 (Proxy) QNN_DLC 18.167 ms 2 - 70 MB NPU BiseNet.dlc
BiseNet w8a8 QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) TFLITE 14.692 ms 8 - 19 MB NPU BiseNet.tflite
BiseNet w8a8 QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) QNN_DLC 10.324 ms 2 - 23 MB NPU BiseNet.dlc
BiseNet w8a8 QCS8550 (Proxy) Qualcomm® QCS8550 (Proxy) ONNX 50.489 ms 63 - 168 MB NPU BiseNet.onnx.zip
BiseNet w8a8 QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) TFLITE 15.014 ms 8 - 46 MB NPU BiseNet.tflite
BiseNet w8a8 QCS9075 (Proxy) Qualcomm® QCS9075 (Proxy) QNN_DLC 11.014 ms 2 - 53 MB NPU BiseNet.dlc
BiseNet w8a8 RB3 Gen 2 (Proxy) Qualcomm® QCS6490 (Proxy) TFLITE 54.143 ms 8 - 66 MB NPU BiseNet.tflite
BiseNet w8a8 RB3 Gen 2 (Proxy) Qualcomm® QCS6490 (Proxy) QNN_DLC 56.611 ms 2 - 85 MB NPU BiseNet.dlc
BiseNet w8a8 RB3 Gen 2 (Proxy) Qualcomm® QCS6490 (Proxy) ONNX 523.392 ms 214 - 227 MB CPU BiseNet.onnx.zip
BiseNet w8a8 RB5 (Proxy) Qualcomm® QCS8250 (Proxy) TFLITE 290.627 ms 8 - 10 MB NPU BiseNet.tflite
BiseNet w8a8 RB5 (Proxy) Qualcomm® QCS8250 (Proxy) ONNX 499.373 ms 185 - 208 MB CPU BiseNet.onnx.zip
BiseNet w8a8 SA7255P ADP Qualcomm® SA7255P TFLITE 24.376 ms 8 - 46 MB NPU BiseNet.tflite
BiseNet w8a8 SA7255P ADP Qualcomm® SA7255P QNN_DLC 21.397 ms 2 - 51 MB NPU BiseNet.dlc
BiseNet w8a8 SA8255 (Proxy) Qualcomm® SA8255P (Proxy) TFLITE 15.026 ms 8 - 19 MB NPU BiseNet.tflite
BiseNet w8a8 SA8255 (Proxy) Qualcomm® SA8255P (Proxy) QNN_DLC 10.338 ms 2 - 21 MB NPU BiseNet.dlc
BiseNet w8a8 SA8295P ADP Qualcomm® SA8295P TFLITE 17.545 ms 8 - 50 MB NPU BiseNet.tflite
BiseNet w8a8 SA8295P ADP Qualcomm® SA8295P QNN_DLC 13.65 ms 2 - 57 MB NPU BiseNet.dlc
BiseNet w8a8 SA8650 (Proxy) Qualcomm® SA8650P (Proxy) TFLITE 14.977 ms 8 - 19 MB NPU BiseNet.tflite
BiseNet w8a8 SA8650 (Proxy) Qualcomm® SA8650P (Proxy) QNN_DLC 10.385 ms 2 - 19 MB NPU BiseNet.dlc
BiseNet w8a8 SA8775P ADP Qualcomm® SA8775P TFLITE 15.014 ms 8 - 46 MB NPU BiseNet.tflite
BiseNet w8a8 SA8775P ADP Qualcomm® SA8775P QNN_DLC 11.014 ms 2 - 53 MB NPU BiseNet.dlc
BiseNet w8a8 Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile TFLITE 10.868 ms 6 - 66 MB NPU BiseNet.tflite
BiseNet w8a8 Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile QNN_DLC 7.35 ms 2 - 72 MB NPU BiseNet.dlc
BiseNet w8a8 Samsung Galaxy S24 Snapdragon® 8 Gen 3 Mobile ONNX 37.823 ms 54 - 1421 MB NPU BiseNet.onnx.zip
BiseNet w8a8 Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile TFLITE 8.25 ms 7 - 48 MB NPU BiseNet.tflite
BiseNet w8a8 Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile QNN_DLC 5.682 ms 2 - 61 MB NPU BiseNet.dlc
BiseNet w8a8 Samsung Galaxy S25 Snapdragon® 8 Elite For Galaxy Mobile ONNX 29.961 ms 69 - 645 MB NPU BiseNet.onnx.zip
BiseNet w8a8 Snapdragon X Elite CRD Snapdragon® X Elite QNN_DLC 10.898 ms 1 - 1 MB NPU BiseNet.dlc
BiseNet w8a8 Snapdragon X Elite CRD Snapdragon® X Elite ONNX 48.511 ms 58 - 58 MB NPU BiseNet.onnx.zip

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

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 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. 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 --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.bisenet.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 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
112
Inference Providers NEW
Image Segmentation
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support