界面已本地化;技能介绍与说明文档按所选语言由 AI 自动翻译,代码、命令、路径和专业术语保留原文,以原作者文档为准。
探索技能
数据与研究 · SKILL.md

data-and-configs

data-and-configs

【数据与研究】Acquire MonoGS datasets, validate dataset layouts, and edit inherited YAML configs for TUM, Replica, EuRoC, and RealSense workflows.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准

技能说明

GitHub

name: data-and-configs description: "Acquire MonoGS datasets, validate dataset layouts, and edit inherited YAML configs for TUM, Replica, EuRoC, and RealSense workflows." disable-model-invocation: true metadata: disco-role: operating license: NOASSERTION

Data and Configs

Use this sub-skill to prepare dataset roots, inspect config inheritance, and preflight MonoGS YAML files before a run.

Start here

Covers

  • recursive inherit_from config merging
  • safe edits to scene-level YAML overrides
  • TUM, Replica, EuRoC, and RealSense dataset roots
  • dataset layout validation with --check-files
  • safe dataset download helpers with a configurable target root

Does not cover

  • CUDA or compiler installation
  • SLAM execution or runtime tuning
  • evaluation metrics or result interpretation
  • live camera GUI operation

Typical use

  • Validate a config and its dataset root: python scripts/validate_monogs_config.py --check-files configs/mono/tum/fr3_office.yaml
  • Download the bundled sample datasets: bash scripts/download_tum.sh --target-root datasets
我的收藏