---
name: yolov8-custom-classification
description: Use for Seeed Jetson topic "Train and deploy a custom classification model with YOLOv8". Includes local references and execution-safe workflow with prerequisites, commands, and validation checks.
---

# Train and deploy a custom classification model with YOLOv8

## Execution model
1. Confirm board model and JetPack/L4T version before applying commands.
2. Read `references/source.body.md` first to extract prerequisites and command order.
3. Execute steps in smallest reproducible sequence and record command outputs.
4. Validate final expected results and keep rollback notes.

## Source
- Original markdown: `/home/darklee/tmp/jetson-develop/refer-参考完后会清除/Application/Computer_Vision/YOLOv8_custom_classification_model.md`
- Scope: `application` (Computer_Vision)

## Topic summary
On this guide we will explain how to train and deploy a custom classification model with YOLOv8

## Key sections
- Introduction
- Overview
- Materials Requirements
- Preparing the reComputer
- Training the model
- Creating the environment
- Exporting the model
- Inference
- ✨ Contributor Project
- Tech Support & Product Discussion

## Reference files
- `references/source.md`
- `references/source.body.md`

## Agent instruction
- Do not skip prerequisite checks.
- If command output differs from expected behavior, pause and ask user before destructive steps.
- Prefer reversible changes and include verification commands.