README.md

---
library_name: pytorch
license: other
tags:
- real_time
- android
pipeline_tag: object-detection

---

![](https://qaihub-public-assets.s3.us-west-2.amazonaws.com/qai-hub-models/models/yolov8_det/web-assets/model_demo.png)

# 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](https://github.com/ultralytics/ultralytics/tree/main/ultralytics/models/yolo/detect).


This repository provides scripts to run YOLOv8-Detection on Qualcomm® devices.
More details on model performance across various devices, can be found
[here](https://aihub.qualcomm.com/models/yolov8_det).

**WARNING**: The model assets are not readily available for download due to licensing restrictions.

### Model Details

- **Model Type:** Model_use_case.object_detection
- **Model Stats:**
  - Model checkpoint: YOLOv8-N
  - Input resolution: 640x640
  - Number of parameters: 3.18M
  - Model size (float): 12.2 MB
  - Model size (w8a8): 3.25 MB
  - Model size (w8a16): 3.60 MB

| Model | Precision | Device | Chipset | Target Runtime | Inference Time (ms) | Peak Memory Range (MB) | Primary Compute Unit | Target Model
|---|---|---|---|---|---|---|---|---|
| YOLOv8-Detection | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 13.032 ms | 0 - 227 MB | NPU | -- |
| YOLOv8-Detection | float | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 13.001 ms | 3 - 233 MB | NPU | -- |
| YOLOv8-Detection | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 6.852 ms | 0 - 173 MB | NPU | -- |
| YOLOv8-Detection | float | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 6.899 ms | 5 - 175 MB | NPU | -- |
| YOLOv8-Detection | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 3.356 ms | 0 - 3 MB | NPU | -- |
| YOLOv8-Detection | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 3.282 ms | 3 - 5 MB | NPU | -- |
| YOLOv8-Detection | float | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 5.443 ms | 5 - 11 MB | NPU | -- |
| YOLOv8-Detection | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 5.02 ms | 0 - 200 MB | NPU | -- |
| YOLOv8-Detection | float | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 4.95 ms | 1 - 214 MB | NPU | -- |
| YOLOv8-Detection | float | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 13.032 ms | 0 - 227 MB | NPU | -- |
| YOLOv8-Detection | float | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 13.001 ms | 3 - 233 MB | NPU | -- |
| YOLOv8-Detection | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 3.358 ms | 0 - 3 MB | NPU | -- |
| YOLOv8-Detection | float | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 3.311 ms | 1 - 3 MB | NPU | -- |
| YOLOv8-Detection | float | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 7.09 ms | 0 - 151 MB | NPU | -- |
| YOLOv8-Detection | float | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 7.139 ms | 0 - 147 MB | NPU | -- |
| YOLOv8-Detection | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 3.35 ms | 0 - 3 MB | NPU | -- |
| YOLOv8-Detection | float | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 3.314 ms | 3 - 5 MB | NPU | -- |
| YOLOv8-Detection | float | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 5.02 ms | 0 - 200 MB | NPU | -- |
| YOLOv8-Detection | float | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 4.95 ms | 1 - 214 MB | NPU | -- |
| YOLOv8-Detection | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 2.435 ms | 0 - 392 MB | NPU | -- |
| YOLOv8-Detection | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 2.473 ms | 4 - 391 MB | NPU | -- |
| YOLOv8-Detection | float | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 3.459 ms | 3 - 210 MB | NPU | -- |
| YOLOv8-Detection | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 1.896 ms | 0 - 225 MB | NPU | -- |
| YOLOv8-Detection | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 1.863 ms | 5 - 213 MB | NPU | -- |
| YOLOv8-Detection | float | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 2.905 ms | 1 - 166 MB | NPU | -- |
| YOLOv8-Detection | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 1.434 ms | 0 - 227 MB | NPU | -- |
| YOLOv8-Detection | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 1.463 ms | 4 - 235 MB | NPU | -- |
| YOLOv8-Detection | float | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 2.424 ms | 1 - 151 MB | NPU | -- |
| YOLOv8-Detection | float | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 3.735 ms | 5 - 5 MB | NPU | -- |
| YOLOv8-Detection | float | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 5.654 ms | 5 - 5 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | QNN_DLC | 20.135 ms | 2 - 151 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | ONNX | 168.676 ms | 65 - 80 MB | CPU | -- |
| YOLOv8-Detection | w8a16 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | QNN_DLC | 8.823 ms | 2 - 6 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | ONNX | 334.433 ms | 63 - 69 MB | CPU | -- |
| YOLOv8-Detection | w8a16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 6.543 ms | 1 - 143 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 3.955 ms | 2 - 173 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 3.269 ms | 2 - 5 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 4.848 ms | 2 - 6 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 16.107 ms | 0 - 143 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 167.306 ms | 59 - 64 MB | CPU | -- |
| YOLOv8-Detection | w8a16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 6.543 ms | 1 - 143 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 3.257 ms | 2 - 4 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 4.501 ms | 0 - 149 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 3.255 ms | 2 - 4 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 16.107 ms | 0 - 143 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 2.198 ms | 0 - 170 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 2.889 ms | 0 - 158 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 1.513 ms | 0 - 148 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 2.275 ms | 0 - 133 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | QNN_DLC | 3.846 ms | 2 - 153 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | ONNX | 158.455 ms | 70 - 87 MB | CPU | -- |
| YOLOv8-Detection | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 1.268 ms | 2 - 152 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 2.075 ms | 0 - 136 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 3.612 ms | 2 - 2 MB | NPU | -- |
| YOLOv8-Detection | w8a16 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 4.944 ms | 2 - 2 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | TFLITE | 8.619 ms | 0 - 137 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | QNN_DLC | 8.474 ms | 1 - 136 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | ONNX | 42.014 ms | 21 - 37 MB | CPU | -- |
| YOLOv8-Detection | w8a8 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | TFLITE | 3.621 ms | 0 - 7 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | QNN_DLC | 3.575 ms | 0 - 3 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Dragonwing RB3 Gen 2 Vision Kit | Qualcomm® QCS6490 | ONNX | 58.059 ms | 21 - 31 MB | CPU | -- |
| YOLOv8-Detection | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | TFLITE | 3.188 ms | 0 - 129 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 3.059 ms | 1 - 130 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | TFLITE | 1.662 ms | 0 - 152 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | QCS8450 (Proxy) | Qualcomm® QCS8450 (Proxy) | QNN_DLC | 1.614 ms | 1 - 151 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | TFLITE | 1.351 ms | 0 - 3 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 1.342 ms | 1 - 3 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | ONNX | 1.814 ms | 0 - 5 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | TFLITE | 1.792 ms | 0 - 129 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 1.761 ms | 1 - 130 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | TFLITE | 35.455 ms | 6 - 28 MB | GPU | -- |
| YOLOv8-Detection | w8a8 | RB5 (Proxy) | Qualcomm® QCS8250 (Proxy) | ONNX | 38.32 ms | 15 - 21 MB | CPU | -- |
| YOLOv8-Detection | w8a8 | SA7255P ADP | Qualcomm® SA7255P | TFLITE | 3.188 ms | 0 - 129 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 3.059 ms | 1 - 130 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | TFLITE | 1.347 ms | 0 - 2 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 1.343 ms | 1 - 3 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | SA8295P ADP | Qualcomm® SA8295P | TFLITE | 2.194 ms | 0 - 136 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | SA8295P ADP | Qualcomm® SA8295P | QNN_DLC | 2.159 ms | 0 - 135 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | TFLITE | 1.364 ms | 0 - 3 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 1.347 ms | 1 - 3 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | SA8775P ADP | Qualcomm® SA8775P | TFLITE | 1.792 ms | 0 - 129 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 1.761 ms | 1 - 130 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | TFLITE | 0.93 ms | 0 - 151 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 0.919 ms | 1 - 152 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | ONNX | 1.212 ms | 0 - 139 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | TFLITE | 0.689 ms | 0 - 134 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 0.665 ms | 1 - 135 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | ONNX | 0.924 ms | 0 - 117 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | TFLITE | 1.526 ms | 0 - 131 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | QNN_DLC | 1.482 ms | 1 - 136 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | ONNX | 39.254 ms | 21 - 40 MB | CPU | -- |
| YOLOv8-Detection | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | TFLITE | 0.626 ms | 0 - 136 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 0.606 ms | 1 - 136 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | ONNX | 0.846 ms | 0 - 134 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 1.558 ms | 1 - 1 MB | NPU | -- |
| YOLOv8-Detection | w8a8 | Snapdragon X Elite CRD | Snapdragon® X Elite | ONNX | 1.77 ms | 2 - 2 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | Dragonwing Q-6690 MTP | Qualcomm® Qcm6690 | QNN_DLC | 12.183 ms | 2 - 146 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | QCS8275 (Proxy) | Qualcomm® QCS8275 (Proxy) | QNN_DLC | 4.392 ms | 1 - 135 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | QCS8550 (Proxy) | Qualcomm® QCS8550 (Proxy) | QNN_DLC | 2.126 ms | 2 - 4 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | QCS9075 (Proxy) | Qualcomm® QCS9075 (Proxy) | QNN_DLC | 2.66 ms | 1 - 135 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | SA7255P ADP | Qualcomm® SA7255P | QNN_DLC | 4.392 ms | 1 - 135 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | SA8255 (Proxy) | Qualcomm® SA8255P (Proxy) | QNN_DLC | 2.123 ms | 4 - 6 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | SA8650 (Proxy) | Qualcomm® SA8650P (Proxy) | QNN_DLC | 2.131 ms | 2 - 4 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | SA8775P ADP | Qualcomm® SA8775P | QNN_DLC | 2.66 ms | 1 - 135 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | Samsung Galaxy S24 | Snapdragon® 8 Gen 3 Mobile | QNN_DLC | 1.422 ms | 2 - 161 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | Samsung Galaxy S25 | Snapdragon® 8 Elite For Galaxy Mobile | QNN_DLC | 1.019 ms | 2 - 139 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | Snapdragon 7 Gen 4 QRD | Snapdragon® 7 Gen 4 Mobile | QNN_DLC | 2.536 ms | 2 - 140 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | Snapdragon 8 Elite Gen 5 QRD | Snapdragon® 8 Elite Gen5 Mobile | QNN_DLC | 0.852 ms | 2 - 141 MB | NPU | -- |
| YOLOv8-Detection | w8a8_mixed_int16 | Snapdragon X Elite CRD | Snapdragon® X Elite | QNN_DLC | 2.424 ms | 2 - 2 MB | NPU | -- |




## Installation


Install the package via pip:
```bash
# NOTE: 3.10 <= PYTHON_VERSION < 3.14 is supported.
pip install "qai-hub-models[yolov8-det]"
```


## Configure Qualcomm® AI Hub Workbench to run this model on a cloud-hosted device

Sign-in to [Qualcomm® AI Hub Workbench](https://workbench.aihub.qualcomm.com/) 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.
```bash
qai-hub configure --api_token API_TOKEN
```
Navigate to [docs](https://workbench.aihub.qualcomm.com/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.

```bash
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.

```bash
python -m qai_hub_models.models.yolov8_det.export
```



## How does this work?

This [export script](https://aihub.qualcomm.com/models/yolov8_det/qai_hub_models/models/YOLOv8-Detection/export.py)
leverages [Qualcomm® AI Hub](https://aihub.qualcomm.com/) 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.

```python
import torch

import qai_hub as hub
from qai_hub_models.models.yolov8_det 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.
```python
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.
```python
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 Workbench. [Sign up for access](https://myaccount.qualcomm.com/signup).



## Run demo on a cloud-hosted device

You can also run the demo on-device.

```bash
python -m qai_hub_models.models.yolov8_det.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.yolov8_det.demo -- --eval-mode on-device
```


## Deploying compiled model to Android


The models can be deployed using multiple runtimes:
- TensorFlow Lite (`.tflite` export): [This
  tutorial](https://www.tensorflow.org/lite/android/quickstart) provides a
  guide to deploy the .tflite model in an Android application.


- QNN (`.so` export ): This [sample
  app](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/sample_app.html)
provides instructions on how to use the `.so` shared library  in an Android application.


## View on Qualcomm® AI Hub
Get more details on YOLOv8-Detection's performance across various devices [here](https://aihub.qualcomm.com/models/yolov8_det).
Explore all available models on [Qualcomm® AI Hub](https://aihub.qualcomm.com/)


## License
* The license for the original implementation of YOLOv8-Detection can be found
  [here](https://github.com/ultralytics/ultralytics/blob/main/LICENSE).



## References
* [Ultralytics YOLOv8 Docs: Object Detection](https://docs.ultralytics.com/tasks/detect/)
* [Source Model Implementation](https://github.com/ultralytics/ultralytics/tree/main/ultralytics/models/yolo/detect)



## Community
* Join [our AI Hub Slack community](https://aihub.qualcomm.com/community/slack) to collaborate, post questions and learn more about on-device AI.
* For questions or feedback please [reach out to us](mailto:ai-hub-support@qti.qualcomm.com).