model-train-infer-backtest-report SKILL.md skillA
model-train-infer-backtest-report from qusong0627/quantmind is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add qusong0627/quantmind/model-train-infer-backtest-report@git:20260919.a5c8006Pins git:20260919.a5c8006 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- qusong0627/quantmind
- Path
skills/model-train-infer-backtest-report/SKILL.md- Versions recorded
- 3
- Latest version
- git:20260919.a5c8006, committed Sep 19, 2026
- First version recorded
- Aug 29, 2026
- Size
- 14,894 bytes
- Repository stars
- 1,774
- License
- AGPL-3.0
模型训练-推理-组合回测-专业报告 全流程 — 提交T+N周期模型训练（13种模型类型：lightgbm/xgboost/catboost/random_forest/linear/mlp/gru/lstm/alstm/transformer/tabnet/tcn/nativetft，GPU训练+GPU推理）、批量推理全年、自定义组合策略回测(分数阈值+大盘MA+止损)、导出研报级MD+PDF报告。在 QuantBot / Claude Code 中训练模型、推理历史、回测策略、对比T+N周期、止损、大盘过滤、出报告、效益分析时使用。触发词：训练模型、模型训练、13种模型、推理全年、回测策略、T+3、T+5、周期对比、止损、大盘过滤、出报告、效益分析.
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
> ⚙️ 本技能遵循公共运行环境契约（最高优先级，先于本文其余内容执行）： > 详见 [_shared/env-contract.md](../_shared/env-contract.md)，执行前先读它。 # 模型训练-推理-回测-报告（全流程）技能 覆盖 QuantMind 量化策略的**完整闭环**：训练模型 → 推理历史 → 组合回测 → 专业报告。用于验证"改个参数（周期/阈值/风控）收益会不会更好"这类问题，用数据说话而不是猜。 ## 架构总览 ``` ① 训练模型 (run-training API) ├─ 模型类型 (13 种全支持): lightgbm / xgboost / catboost / random_forest / | linear / mlp / gru / lstm / alstm / transformer / tabnet / tcn / nativetft ├─ T+N 周期选择 (T+1/T+3/T+5...) ├─ 特征 (75~182 个来自 QuantDB) ├─ GPU: 训练 GPU 满载; 推理已改 CUDA (train.py _predict_dl/_predict_nativetft, 有算力就用 GPU) ├─ 质量门禁: test_rank_icir<0.05 或 test_rank_ic 非正 → status=candidate (产品逻辑, 非失败) | └─ 训练 2023-2025 → 推理 2026 ② 推理历史 (批量推理 API) ├─ 单日 /models/inference/run └─ 批量 range 模式 推理全年逐个交易日 ③ 组合回测 (自定义脚本) ├─ backtest_l2_top20.py 基本 Top20 每日滚动 ├─ backtest_l2_year.py 全年严格版 (滑点+T+1+涨跌停+ST剔除) └─ backtest_l2_optimized.py 优化版 (分数阈值+大盘MA+止损5%) ④ 专业报告 (report_l2_optimized.py) └─ 回测数据 → 研报级 MD → md_to_pdf_report.py → PDF ``` ## 认证 ```bash BASE=http://localhost:8000 TOKEN=$(curl -s -X POST $BASE/api/v1/auth/login -H "Content-Type: application/json" \ …
Read the whole file at its exact version.
How to install
mdr add qusong0627/quantmind/model-train-infer-backtest-report@sha256:1911dec5acd76c21The line at the top of this page pins the label git:20260919.a5c8006, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_zwf2sjj7hwcqwbta)
2 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 3 versions of this file, between Aug 29, 2026 and Sep 19, 2026, about one every 11 days. Its first recorded version is graded B and its latest A.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| git:20260919.a5c8006 latest | 2026-09-19 | a5c8006 | 14,894 B | A | view · diff |
| git:20260829.c9fd96c | 2026-08-29 | c9fd96c | 16,731 B | B | view · diff |
| git:20260829.cda0e48 | 2026-08-29 | cda0e48 | 16,733 B | B | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (14894 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
qusong0627/quantmind · 1,774 stars · license AGPL-3.0 · pushed 2026-10-22 · branch master
API
GET https://markdownregistry.com/api/v1/artifacts/art_zwf2sjj7hwcqwbta GET https://markdownregistry.com/api/v1/resolve?ref=qusong0627/quantmind/model-train-infer-backtest-report GET https://markdownregistry.com/api/v1/blob/1911dec5acd76c217899f1ef04b50d85a114c1eafa23663d60721f47b59a57ed
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.