agentsclimarketplace

Bigdata ml

Skill MARUCIE/openclaw-foundry/web/public/packs/data-analyst/skills/bigdata-ml

Machine learning toolkit for big data teams. Includes scikit-learn, PyTorch Lightning, Transformers, SHAP for model training, deployment, and interpretation. Use when building ML pipelines, training models, or explaining predictions.From its SKILL.md

Install
npx -y skills add MARUCIE/openclaw-foundry --skill bigdata-ml

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.6 KB, ~1.4k tokens by cl100k_base, as published. Nobody here has run it

Big Data Machine Learning Toolkit

Overview

大数据团队机器学习工具集,从传统ML到深度学习全覆盖。

Quick Reference

工具场景规模
scikit-learn传统ML中等数据
PyTorch Lightning深度学习GPU训练
TransformersNLP/LLM预训练模型
SHAP模型解释可解释AI

选择指南

任务类型:
├── 分类/回归 → scikit-learn
├── 时间序列 → scikit-learn + statsmodels
├── 文本处理 → Transformers
├── 图像处理 → PyTorch Lightning
├── 强化学习 → stable-baselines3
└── 模型解释 → SHAP

子Skills

  • scikit-learn/ - 传统机器学习
  • pytorch-lightning/ - 深度学习框架
  • transformers/ - NLP预训练模型
  • shap/ - 模型可解释性
  • stable-baselines3/ - 强化学习

常用模式

标准ML Pipeline (scikit-learn)

from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import cross_val_score

pipeline = Pipeline([
    ('scaler', StandardScaler()),
    ('classifier', RandomForestClassifier())
])

scores = cross_val_score(pipeline, X, y, cv=5)
print(f"CV Score: {scores.mean():.3f} ± {scores.std():.3f}")

深度学习训练 (PyTorch Lightning)

import pytorch_lightning as pl
from pytorch_lightning.callbacks import EarlyStopping

trainer = pl.Trainer(
    max_epochs=100,
    accelerator="gpu",
    callbacks=[EarlyStopping(monitor="val_loss")]
)
trainer.fit(model, train_loader, val_loader)

NLP任务 (Transformers)

from transformers import pipeline

# 文本分类
classifier = pipeline("text-classification", model="bert-base-chinese")
result = classifier("这个产品质量很好")

# 文本生成
generator = pipeline("text-generation", model="gpt2")

模型解释 (SHAP)

import shap

explainer = shap.TreeExplainer(model)
shap_values = explainer.shap_values(X)

# 特征重要性
shap.summary_plot(shap_values, X)

# 单样本解释
shap.force_plot(explainer.expected_value, shap_values[0], X.iloc[0])

大数据ML最佳实践

1. 大规模训练

# 使用Dask-ML
from dask_ml.model_selection import train_test_split
from dask_ml.linear_model import LogisticRegression

X_train, X_test = train_test_split(X_dask, y_dask)
model = LogisticRegression()
model.fit(X_train, y_train)

2. 增量学习

from sklearn.linear_model import SGDClassifier

model = SGDClassifier()
for chunk in data_chunks:
    model.partial_fit(chunk.X, chunk.y, classes=[0, 1])

3. 模型版本管理

import mlflow

with mlflow.start_run():
    mlflow.log_params(params)
    mlflow.log_metrics(metrics)
    mlflow.sklearn.log_model(model, "model")

团队规范

  1. 实验追踪: 使用MLflow记录所有实验
  2. 模型版本: 模型+数据版本绑定
  3. 可解释性: 关键模型必须提供SHAP解释
  4. 部署流程: 模型 → 测试 → 审核 → 上线

猪哥云-数据产品部 | 大数据团队专用

是什么

Big Data Machine Learning Toolkit 用来把 数据分析师 场景里的任务输入转成可执行的流程、检查清单和交付物。

Machine learning toolkit for big data teams. Includes scikit-learn, PyTorch Lightning, Transformers, SHAP for model training, deployment, and interpretation. Use when building ML pipelines, training models, or explaining predictions.

它的价值在于让 数据AI职能线 在 Claude Code、Codex、Gemini、Hermes 或 OpenClaw 中复用同一套岗位能力,而不是依赖一次性的聊天提示词。

怎么用

  1. 明确当前任务目标、输入材料、约束和期望交付物,再加载 bigdata-ml
  2. 按 skill 文档中的流程、检查清单或工具建议执行,优先复用仓库已有规范与真实命令。
  3. 把关键判断、风险、验证命令和产出路径记录到当前任务文档或交付说明中。
  4. 用最小可证明的检查确认结果有效;发现缺口时回到 skill 清单补齐。

架构图

flowchart LR
  A[任务输入] --> B[加载 Big Data Machine Learning Toolkit]
  B --> C[执行流程与检查清单]
  C --> D[生成交付物与风险记录]
  D --> E[验证结果并沉淀复盘]

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.