代码仓库
评测基准 / 科研
faultline · amanshah0729
Faultline使用扰动和消融工作流程来研究LIBERO策略稳健性。
概述
Faultline studies LIBERO policy robustness using perturbation and ablation workflows.<br><br>Publisher summary：“Measuring what a LIBERO success rate hides. Perturbation + ablation harness, 4 policies, ~2,000 rollouts, one laptop, no CUDA.”<br><br>Use its repository to inspect the implementation, documentation and supported workflows. This directory record identifies the maintainer, code license and related resources; it does not claim that the project has been installed or tested on your robot.
项目信息
- 维护者
- amanshah0729
- 主要语言
- Python
- 存储库代码许可证
- 未指定
- 模型重量许可证
- 未记录
- 数据集许可
- 未记录
- 硬件设计许可证
- 未记录
- 维护
- 未存档在最后记录的元数据中
- 最后观察到的存储库推送
- 2026-08-14
代码许可不会自动涵盖模型权重、数据集或硬件设计。未记录的许可证详细信息需要与发布者确认。
文档部分
- Faultline
- Why this matters
- Result 1 — the benchmark cannot distinguish policies that are not close
- Result 2 — these policies ignore inputs they are assumed to use
- Result 3 — the most damaging perturbation is the worst one for comparing policies
- What is and isn't new here
浏览该资源的类别
探索相关集合
浏览 评测基准 →更多内容见 评测基准
代码仓库
