AgentSkillsCN

train

使用 Ultralytics 在已标注的数据集上训练 YOLO 模型。自动将数据划分为训练集与验证集，生成 dataset.yaml 文件，并输出训练好的权重。建议在完成标注与数据增强后使用。

SKILL.md
--- frontmatter
name: train
description: Train a YOLO model on labeled dataset using ultralytics. Splits data into train/val, generates dataset.yaml, and produces trained weights. Use after labeling and augmenting.

Instructions

  1. •Read config.json for yolo_model, epochs, train_split, output_dir
  2. •Run: uv run .agents/skills/train/scripts/run.py
  3. •Outputs: output/weights/best.pt, output/dataset.yaml