---
license: agpl-3.0
language:
- en
base_model:
- yolov8
pipeline_tag: object-detection
tags:
- Axera
- YOLOv8
- NPU
- Ultralytics
- Object Detection
- Aquarium
---

# YOLOv8-Aquarium

This version of **YOLOv8-Aquarium** has been converted to run on the Axera NPU using **w8a16** quantization. It is fine-tuned to detect aquatic life in aquarium environments.

## Supported Classes

This model is trained to detect the following **5 classes**:
1. **fish**
2. **turtle**
3. **shrimp**
4. **crab**
5. **snail**

Compatible with Pulsar2 version: 5.1.

## Convert tools links:

For those who are interested in model conversion, you can try to export axmodel through:
- [The repo of AXera Platform](https://github.com/AXERA-TECH/ax-samples), where you can get the detailed guide.
- [Pulsar2 Link, How to Convert ONNX to axmodel](https://pulsar2-docs.readthedocs.io/en/latest/pulsar2/introduction.html)

## Support Platform

https://docs.m5stack.com/zh_CN/ai_hardware/AI_Pyramid-Pro

- **AX650N/AX8850**
  - [M4N-Dock(爱芯派Pro)](https://wiki.sipeed.com/hardware/zh/maixIV/m4ndock/m4ndock.html)
  - [AI Pyramid](https://docs.m5stack.com/zh_CN/ai_hardware/AI_Pyramid-Pro)
  - [M.2 Accelerator card](https://docs.m5stack.com/en/ai_hardware/LLM-8850_Card)
- **AX620E**
  - [爱芯派2](https://axera-pi-2-docs-cn.readthedocs.io/zh-cn/latest/index.html)
  - [Module-LLM](https://docs.m5stack.com/zh_CN/module/Module-LLM)
- **AX637**

## How to use

Download all files from this repository to the device.

### python env requirement

#### pyaxengine

https://github.com/AXERA-TECH/pyaxengine

```bash
wget https://github.com/AXERA-TECH/pyaxengine/releases/download/0.1.3.rc2/axengine-0.1.3-py3-none-any.whl
pip install axengine-0.1.3-py3-none-any.whl
```

### Inference with AX650 Host, such as M4N-Dock(爱芯派Pro)

Input image:
![](test.png)

run
```bash
python3 infer_yolov8_pyax.py --model-path aquarium_yolov8s_6way.axmodel --test-img test.png
```

```bash
root@ax650:~/yolo8# python3 infer_yolov8_pyax.py --model-path aquarium_yolov8s_6way.axmodel --test-img test.png
[INFO] Available providers:  ['AxEngineExecutionProvider', 'AXCLRTExecutionProvider']
[INFO] Using provider: AxEngineExecutionProvider
[INFO] Chip type: ChipType.MC50
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Engine version: 2.12.0s
[INFO] Model type: 2 (triple core)
[INFO] Compiler version: 5.2-dirty a6f1799a-dirty
[Aquarium-YOLOv8-6way] [14:15:20.397] [DEBUG] Load model time = 485.48 ms
[Aquarium-YOLOv8-6way] [14:15:20.463] [DEBUG] Pre-process time = 8.36 ms
[Aquarium-YOLOv8-6way] [14:15:20.483] [DEBUG] Forward time min/avg/max = 19.42/19.42/19.42 ms (n=1)
[Aquarium-YOLOv8-6way] [14:15:20.490] [DEBUG] Post-process time = 6.78 ms
[Aquarium-YOLOv8-6way] [14:15:20.491] [INFO] Detections: 26
[Aquarium-YOLOv8-6way] [14:15:20.491] [INFO]   fish     score=0.850  xyxy=(168.9,517.7,285.1,583.1)
[Aquarium-YOLOv8-6way] [14:15:20.491] [INFO]   fish     score=0.771  xyxy=(757.8,573.3,801.3,601.8)
...
[Aquarium-YOLOv8-6way] [14:15:20.496] [INFO]   fish     score=0.452  xyxy=(600.2,412.4,652.9,460.7)
[Aquarium-YOLOv8-6way] [14:15:20.496] [INFO] per-class: {'fish': 26, 'turtle': 0, 'shrimp': 0, 'crab': 0, 'snail': 0}
[Aquarium-YOLOv8-6way] [14:15:20.526] [INFO] Saved to result_aquarium_yolov8.jpg
root@ax650:~/yolo8# ls
aquarium_yolov8s_6way.axmodel  infer_yolov8_pyax.py  result_aquarium_yolov8.jpg  test.png
root@ax650:~/yolo8# realpath result_aquarium_yolov8.jpg
/root/yolo8/result_aquarium_yolov8.jpg
```

Output image:
![](result_aquarium_yolov8.jpg)
```