Hugging Face's logo

qualcomm
/
YOLOv8-Segmentation

Image Segmentation
PyTorch
TF Lite
ONNX
real_time
android
Model card Files Community
Edit model card

YOLOv8-Segmentation: Optimized for Mobile Deployment

Real-time object segmentation optimized for mobile and edge by Ultralytics

Ultralytics YOLOv8 is a machine learning model that predicts bounding boxes, segmentation masks and classes of objects in an image.

This model is an implementation of YOLOv8-Segmentation found here.

This repository provides scripts to run YOLOv8-Segmentation 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: YOLOv8N-Seg
    • Input resolution: 640x640
    • Number of parameters: 3.43M
    • Model size: 13.2 MB
    • Number of output classes: 80
Model Device Chipset Target Runtime Inference Time (ms) Peak Memory Range (MB) Precision Primary Compute Unit Target Model
YOLOv8-Segmentation Samsung Galaxy S23 Snapdragon® 8 Gen 2 TFLITE 6.58 ms 4 - 7 MB FP16 NPU YOLOv8-Segmentation.tflite
YOLOv8-Segmentation Samsung Galaxy S23 Snapdragon® 8 Gen 2 QNN 6.435 ms 5 - 19 MB FP16 NPU YOLOv8-Segmentation.so
YOLOv8-Segmentation Samsung Galaxy S23 Snapdragon® 8 Gen 2 ONNX 7.637 ms 12 - 20 MB FP16 NPU YOLOv8-Segmentation.onnx
YOLOv8-Segmentation Samsung Galaxy S24 Snapdragon® 8 Gen 3 TFLITE 4.674 ms 4 - 118 MB FP16 NPU YOLOv8-Segmentation.tflite
YOLOv8-Segmentation Samsung Galaxy S24 Snapdragon® 8 Gen 3 QNN 4.601 ms 5 - 61 MB FP16 NPU YOLOv8-Segmentation.so
YOLOv8-Segmentation Samsung Galaxy S24 Snapdragon® 8 Gen 3 ONNX 5.063 ms 16 - 135 MB FP16 NPU YOLOv8-Segmentation.onnx
YOLOv8-Segmentation Snapdragon 8 Elite QRD Snapdragon® 8 Elite TFLITE 4.509 ms 3 - 74 MB FP16 NPU YOLOv8-Segmentation.tflite
YOLOv8-Segmentation Snapdragon 8 Elite QRD Snapdragon® 8 Elite QNN 4.372 ms 5 - 55 MB FP16 NPU Use Export Script
YOLOv8-Segmentation Snapdragon 8 Elite QRD Snapdragon® 8 Elite ONNX 4.896 ms 0 - 71 MB FP16 NPU YOLOv8-Segmentation.onnx
YOLOv8-Segmentation QCS8550 (Proxy) QCS8550 Proxy TFLITE 6.559 ms 2 - 10 MB FP16 NPU YOLOv8-Segmentation.tflite
YOLOv8-Segmentation QCS8550 (Proxy) QCS8550 Proxy QNN 6.321 ms 5 - 9 MB FP16 NPU Use Export Script
YOLOv8-Segmentation SA8255 (Proxy) SA8255P Proxy TFLITE 6.473 ms 0 - 175 MB FP16 NPU YOLOv8-Segmentation.tflite
YOLOv8-Segmentation SA8255 (Proxy) SA8255P Proxy QNN 6.295 ms 5 - 11 MB FP16 NPU Use Export Script
YOLOv8-Segmentation SA8775 (Proxy) SA8775P Proxy TFLITE 6.453 ms 4 - 7 MB FP16 NPU YOLOv8-Segmentation.tflite
YOLOv8-Segmentation SA8775 (Proxy) SA8775P Proxy QNN 6.381 ms 5 - 6 MB FP16 NPU Use Export Script
YOLOv8-Segmentation SA8650 (Proxy) SA8650P Proxy TFLITE 6.424 ms 1 - 3 MB FP16 NPU YOLOv8-Segmentation.tflite
YOLOv8-Segmentation SA8650 (Proxy) SA8650P Proxy QNN 6.278 ms 5 - 6 MB FP16 NPU Use Export Script
YOLOv8-Segmentation SA8295P ADP SA8295P TFLITE 10.892 ms 4 - 65 MB FP16 NPU YOLOv8-Segmentation.tflite
YOLOv8-Segmentation SA8295P ADP SA8295P QNN 10.437 ms 0 - 5 MB FP16 NPU Use Export Script
YOLOv8-Segmentation QCS8450 (Proxy) QCS8450 Proxy TFLITE 9.546 ms 4 - 103 MB FP16 NPU YOLOv8-Segmentation.tflite
YOLOv8-Segmentation QCS8450 (Proxy) QCS8450 Proxy QNN 9.162 ms 5 - 45 MB FP16 NPU Use Export Script
YOLOv8-Segmentation Snapdragon X Elite CRD Snapdragon® X Elite QNN 6.852 ms 5 - 5 MB FP16 NPU Use Export Script
YOLOv8-Segmentation Snapdragon X Elite CRD Snapdragon® X Elite ONNX 7.643 ms 17 - 17 MB FP16 NPU YOLOv8-Segmentation.onnx

Installation

This model can be installed as a Python package via pip.

pip install "qai-hub-models[yolov8_seg]"

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.yolov8_seg.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.yolov8_seg.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.yolov8_seg.export

Profiling Results
------------------------------------------------------------
YOLOv8-Segmentation
Device                          : Samsung Galaxy S23 (13)
Runtime                         : TFLITE                 
Estimated inference time (ms)   : 6.6                    
Estimated peak memory usage (MB): [4, 7]                 
Total # Ops                     : 338                    
Compute Unit(s)                 : NPU (338 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.yolov8_seg import 

# Load the model

# Device
device = hub.Device("Samsung Galaxy S23")

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.yolov8_seg.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.yolov8_seg.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 YOLOv8-Segmentation's performance across various devices here. Explore all available models on Qualcomm® AI Hub

License

  • The license for the original implementation of YOLOv8-Segmentation 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 Examples
Image Segmentation
Inference API (serverless) does not yet support pytorch models for this pipeline type.