YOLOv8-Detection: Optimized for Mobile Deployment
Real-time object detection optimized for mobile and edge by Ultralytics
Ultralytics YOLOv8 is a machine learning model that predicts bounding boxes and classes of objects in an image.
This model is an implementation of YOLOv8-Detection found here.
This repository provides scripts to run YOLOv8-Detection on Qualcomm® devices. More details on model performance across various devices, can be found here.
Model Details
- Model Type: Object detection
- Model Stats:
- Model checkpoint: YOLOv8-N
- Input resolution: 640x640
- Number of parameters: 3.18M
- Model size: 12.2 MB
| Model | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Precision | Primary Compute Unit | Target Model |
|---|---|---|---|---|---|---|---|---|
| YOLOv8-Detection | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | TFLITE | 5.164 ms | 0 - 4 MB | FP16 | NPU | YOLOv8-Detection.tflite |
| YOLOv8-Detection | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | QNN | 5.216 ms | 5 - 20 MB | FP16 | NPU | YOLOv8-Detection.so |
| YOLOv8-Detection | Samsung Galaxy S23 | Snapdragon® 8 Gen 2 | ONNX | 6.065 ms | 5 - 9 MB | FP16 | NPU | YOLOv8-Detection.onnx |
| YOLOv8-Detection | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | TFLITE | 3.722 ms | 0 - 94 MB | FP16 | NPU | YOLOv8-Detection.tflite |
| YOLOv8-Detection | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | QNN | 3.708 ms | 5 - 56 MB | FP16 | NPU | YOLOv8-Detection.so |
| YOLOv8-Detection | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 | ONNX | 4.087 ms | 5 - 117 MB | FP16 | NPU | YOLOv8-Detection.onnx |
| YOLOv8-Detection | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | TFLITE | 3.674 ms | 0 - 59 MB | FP16 | NPU | YOLOv8-Detection.tflite |
| YOLOv8-Detection | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | QNN | 3.606 ms | 5 - 48 MB | FP16 | NPU | Use Export Script |
| YOLOv8-Detection | Snapdragon 8 Elite QRD | Snapdragon® 8 Elite | ONNX | 3.977 ms | 5 - 72 MB | FP16 | NPU | YOLOv8-Detection.onnx |
| YOLOv8-Detection | QCS8550 (Proxy) | QCS8550 Proxy | TFLITE | 5.146 ms | 5 - 6 MB | FP16 | NPU | YOLOv8-Detection.tflite |
| YOLOv8-Detection | QCS8550 (Proxy) | QCS8550 Proxy | QNN | 4.992 ms | 5 - 6 MB | FP16 | NPU | Use Export Script |
| YOLOv8-Detection | SA8255 (Proxy) | SA8255P Proxy | TFLITE | 5.177 ms | 0 - 5 MB | FP16 | NPU | YOLOv8-Detection.tflite |
| YOLOv8-Detection | SA8255 (Proxy) | SA8255P Proxy | QNN | 5.011 ms | 6 - 7 MB | FP16 | NPU | Use Export Script |
| YOLOv8-Detection | SA8775 (Proxy) | SA8775P Proxy | TFLITE | 5.171 ms | 0 - 16 MB | FP16 | NPU | YOLOv8-Detection.tflite |
| YOLOv8-Detection | SA8775 (Proxy) | SA8775P Proxy | QNN | 5.034 ms | 5 - 6 MB | FP16 | NPU | Use Export Script |
| YOLOv8-Detection | SA8650 (Proxy) | SA8650P Proxy | TFLITE | 5.191 ms | 0 - 2 MB | FP16 | NPU | YOLOv8-Detection.tflite |
| YOLOv8-Detection | SA8650 (Proxy) | SA8650P Proxy | QNN | 5.003 ms | 5 - 6 MB | FP16 | NPU | Use Export Script |
| YOLOv8-Detection | SA8295P ADP | SA8295P | TFLITE | 9.914 ms | 0 - 51 MB | FP16 | NPU | YOLOv8-Detection.tflite |
| YOLOv8-Detection | SA8295P ADP | SA8295P | QNN | 8.548 ms | 0 - 5 MB | FP16 | NPU | Use Export Script |
| YOLOv8-Detection | QCS8450 (Proxy) | QCS8450 Proxy | TFLITE | 8.7 ms | 0 - 83 MB | FP16 | NPU | YOLOv8-Detection.tflite |
| YOLOv8-Detection | QCS8450 (Proxy) | QCS8450 Proxy | QNN | 7.915 ms | 5 - 39 MB | FP16 | NPU | Use Export Script |
| YOLOv8-Detection | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN | 5.532 ms | 5 - 5 MB | FP16 | NPU | Use Export Script |
| YOLOv8-Detection | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 6.7 ms | 5 - 5 MB | FP16 | NPU | YOLOv8-Detection.onnx |
Installation
This model can be installed as a Python package via pip.
pip install "qai-hub-models[yolov8_det]"
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_det.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_det.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_det.export
Profiling Results
------------------------------------------------------------
YOLOv8-Detection
Device : Samsung Galaxy S23 (13)
Runtime : TFLITE
Estimated inference time (ms) : 5.2
Estimated peak memory usage (MB): [0, 4]
Total # Ops : 290
Compute Unit(s) : NPU (290 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_det 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_det.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_det.demo -- --on-device
Deploying compiled model to Android
The models can be deployed using multiple runtimes:
TensorFlow Lite (
.tfliteexport): This tutorial provides a guide to deploy the .tflite model in an Android application.QNN (
.soexport ): This sample app provides instructions on how to use the.soshared library in an Android application.
View on Qualcomm® AI Hub
Get more details on YOLOv8-Detection's performance across various devices here. Explore all available models on Qualcomm® AI Hub
License
- The license for the original implementation of YOLOv8-Detection can be found here.
- The license for the compiled assets for on-device deployment can be found here
References
Community
- Join our AI Hub Slack community to collaborate, post questions and learn more about on-device AI.
- For questions or feedback please reach out to us.