Work with the DOTA research codebase for dataset setup, benchmark execution, configuration changes, reproduction guidance, and troubleshooting. Use when an AI coding agent needs to run or modify DOTA experiments, explain the repository layout, prepare OOD or cross-domain benchmarks, edit dataset/config mappings, inspect logs, or debug DOTA training/evaluation failures.
2026-07-02