feat(training): 实现 LightGBM 模型

- 新增 LightGBMModel:LightGBM 回归模型实现
- 支持自定义参数(objective, num_leaves, learning_rate, n_estimators 等)
- 使用 LightGBM 原生格式保存/加载模型(不依赖 pickle)
- 支持特征重要性提取
- 已注册到 ModelRegistry(@register_model("lightgbm"))
This commit is contained in:
2026-03-03 22:30:37 +08:00
parent 9ca1deae56
commit f35a6a76a6
4 changed files with 432 additions and 0 deletions

View File

@@ -0,0 +1,8 @@
"""模型子模块
包含各种机器学习模型的实现。
"""
from src.training.components.models.lightgbm import LightGBMModel
__all__ = ["LightGBMModel"]