alpharsi
AlphaRSI — three-role self-improvement loop that evolves the training recipe of alpha sequence models. Operator runs walk-forward, Evaluator grades with OOS RankIC, Refiner proposes hyperparameters via TPE Bayesian search; only strictly-better recipes are kept. Offline one-command run, data bundle shipped with sha1 reconciliation, GPU auto-detect. Ships one skill.
AlphaRSI——三角色自改进闭环，进化 alpha 序列模型的训练配方。Operator 跑 walk-forward、Evaluator 用 OOS RankIC 验收、Refiner 用 TPE 贝叶斯搜索提议超参，只保留严格更优。离线一条命令跑，数据包 sha1 对账分发，GPU 自动走 CUDA。附带一个 skill。
How to install
dsh plugin add github:Z-Asset/alpharsi About
AlphaRSI — alpha 模型训练配方自进化 把 alpha 序列模型（如 S13 中证 500 复刻包）的训练配方当进化对象。三角色闭环：Operator 跑 walk-forward、Evaluator 用 OOS_RankIC 验收、Refiner 提议下一组超参，只保留严格更优。参照 MetaRSI-v1 的 Model-RSI 算子与 RSI Kernel 权限边界。 三个角色 | 角色 | 实现 | |---|---| | Operator 运行模型 | train_from_xy.py 的 walk_forward()（确定性训练脚本，非 LLM） | | Evaluator 验收模型 | summarize() 的 OOS_RankIC / ICIR / SR（封闭） | | Refiner 改进模型 | 离线 TPE 贝叶斯搜索；在线 DeepSeek LLM | 循环从不重置：评估 → 诊断 → 提议配方 → 校验 → 评估 → 只保留严格更优 → 继续。 一条命令 bash pip install -e . # 装包（提供 alpharsi 命令） alpharsi <S13数据包路径> # 离线 TPE 搜索，默认参数 alpharsi --doctor <S13数据包路径> # 体检：sha1 对账 + 依赖 + 数据口径 + CUDA alpha…
Recommendation signals
Meta
- License
- –
- Language
- Python
- GitHub stars
- 0
- mo. downloads
- –
- Last push
- 2026-09-17
- Created
- 2026-09-17
Links
Basic safety check
- Findings
- curated 收录但无 npm 包/安装命令；无 license
- Sources
- curated:awesome-dsh-plugin.com, curated:awesome-dsh-plugin/awesome-dsh-plugin
Related plugins
agent-device
callstackincubator/agent-device
Mobile app automation and verification for AI coding agents. CLI, MCP server, and typed Node.js API for iOS, Android, HarmonyOS, TV, web, macOS, and Linux.
SuperClaude_Framework
SuperClaude-Org/SuperClaude_Framework
A configuration framework that enhances Claude Code with specialized commands, cognitive personas, and development methodologies.
pipedream
PipedreamHQ/pipedream
Connect APIs, remarkably fast. Free for developers.