haotian 57d486b0c9 修改--修改文档20250226

2025-02-26 11:07:23 +08:00

3.4 KiB

Raw Blame History

运行说明

1. 启动mlflow

- mlflow server --host 10.0.0.202 --port 5000
- 这里host要和config/config.yaml中配置的host一致, 不然连接不上

2. 运行 main.py

- python main.py
- 运行ip和端口在config/config.yaml中配置
- 当前python环境名为trt
- 在ip:port/docs可以看到可视化接口文档

3. 安装python环境

- 使用conda创建一个新的python环境
- pip 换源 清华源
    -- pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
- pip install -r requirements.txt

4. 测试接口

注: 10.0.0.202为当前本机ip

GET http://10.0.0.202:8992/health

传递参数 { "input_path": "dataset/dataset_raw/breast_cancer.csv", "output_dir": "dataset/dataset_processed", "process_methods": [ { "method_name": "IsolationForest", "params": { "contamination": 0.1, "random_state": 42 } } ], "feature_methods": [

    ],
    "split_params": {
        "test_size": 0.1,
        "val_size": 0.2
    }
}

4.6 获取可用数据集列表

GET http://10.0.0.202:8992/data/datasets

4.7 获取可用模型列表

GET http://10.0.0.202:8992/model/available

4.8 获取模型详情

GET http://10.0.0.202:8992/model/available/{model_name}

4.9 获取评价指标列表

GET http://10.0.0.202:8992/model/metrics

4.10 模型训练

POST http://10.0.0.202:8992/model/train

{ "train_path": "/home/admin-root/haotian/MLPlatform/dataset/dataset_processed/breast_cancer_20250224_170615/train_breast_cancer_20250224_170615.csv", "val_path": "/home/admin-root/haotian/MLPlatform/dataset/dataset_processed/breast_cancer_20250224_170615/val_breast_cancer_20250224_170615.csv", "algorithm": "XGBClassifier", "task_type": "classification", "parameters": { "n_estimators": 100, "learning_rate": 0.1, "max_depth": 6, "random_state": 42 }, "experiment_name": "test_post_1" }

4.11 获取实验列表

GET http://10.0.0.202:8992/model/experiments

4.12 获取具体实验内容

GET http://10.0.0.202:8992/model/experiment/{experiment_name}

4.13 删除模型指定实验模型

DELETE

4.14 模型预测

POST http://10.0.0.202:8992/model/predict { "run_id": "33939ea6d8ce4d43a268f23f7361651e", "data_path": "/home/admin-root/haotian/MLPlatform/dataset/dataset_processed/breast_cancer_20250219_145614/test_breast_cancer_20250219_145614.csv", "output_path": "predictions/pred_breast_cancer_20250219_145614.csv", "return_proba": true, "metrics": [ "accuracy", "f1", "precision", "recall" ] }

4.15 查看系统资源使用情况

GET http://10.0.0.202:8992/system/resources

4.16 查看训练历史

GET http://10.0.0.202:8992/system/history

4.17 查看系统训练日志

GET http://10.0.0.202:8992/system/logs

5. 前端界面

- 暂未完成

3.4 KiB Raw Blame History