harborframework/parity-experiments
收藏资源简介:
# Upload your Adapter Oracle and Parity results This dataset saves the oracle and parity experiment logs for adapters. Please upload them according to the following format and draft a PR. ``` adapters/ └── {adapter_name}/ ├── README.md # Results overview, directory structure, trajectory interpretation, notes, etc. This should be DIFFERENT than the adapter REAMDE. ├── config.yaml # The yaml file that can be directly used to run parity experiments in Harbor. ├── original_parity/ ├── harbor_parity/ ├── oracle/ └── results_collection/ # copy the valid result.json files from parity to this directory ├── result_{original/harbor}_run1.json ├── result_{original/harbor}_run2.json ├── ... └── result_{original/harbor}_run{N}.json ``` We will use them to estimate costs for each adapter on diverse agent and models. ## Git LFS rules Most large binary and media files are already covered by the root `.gitattributes` and need no action. If you do need an extra LFS rule for your adapter, put it in a nested `.gitattributes` inside your adapter directory — **do not modify the repo-root `.gitattributes`**, which is a shared merge-conflict hotspot. Run `git lfs track` from inside your adapter directory so the rule lands in the nested file with a relative pattern: ```bash cd adapters/<adapter_name> git lfs track "path/to/large_file.json" cd - git add adapters/<adapter_name>/.gitattributes ``` --- license: apache-2.0 ---
# 上传你的适配器(Adapter)基准验证与一致性实验结果 本数据集用于存储适配器的基准验证与一致性实验日志,请按照下述格式上传相关文件并提交拉取请求(Pull Request,简称PR)。 adapters/ └── {adapter_name}/ ├── README.md # 实验结果概述、目录结构、轨迹解读、注意事项等,需与适配器自身的README文件有所区分。 ├── config.yaml # 可直接用于在Harbor平台中运行一致性实验的YAML配置文件。 ├── original_parity/ # 原始一致性实验结果目录 ├── harbor_parity/ # Harbor平台一致性实验结果目录 ├── oracle/ # 基准验证实验结果目录 └── results_collection/ # 从一致性实验目录中复制有效result.json文件至此目录 ├── result_{original/harbor}_run1.json # 第1次运行的结果文件 ├── result_{original/harbor}_run2.json # 第2次运行的结果文件 ├── ... └── result_{original/harbor}_run{N}.json # 第N次运行的结果文件 我们将借助这些数据,估算不同智能体(AI Agent)与模型上各适配器的运行成本。 ### Git大文件存储(Git LFS)规则 绝大多数大型二进制与媒体文件已由仓库根目录下的`.gitattributes`文件覆盖,无需额外操作。若需为你的适配器配置专属的LFS规则,请将规则写入适配器目录内的嵌套`.gitattributes`文件中——**请勿修改仓库根目录的`.gitattributes`**,该文件属于共享的合并冲突高发区域。请在适配器目录内执行以下命令,使规则以相对路径格式写入嵌套的`.gitattributes`文件: bash cd adapters/<adapter_name> git lfs track "path/to/large_file.json" cd - git add adapters/<adapter_name>/.gitattributes --- 许可证:Apache-2.0 ---




