# MyMLOPS

**Repository Path**: fubob/my-mlops

## Basic Information

- **Project Name**: MyMLOPS
- **Description**: 用于学习，测试MLOPS相关的工具
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-10-23
- **Last Updated**: 2025-11-25

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# MyMLOPS

## 介绍
用于学习，测试MLOPS相关的工具



## MLflow、Kubeflow 与 Airflow 工具的对比解析

MLflow 轻量化，但是不适用于集群

Kubeflow 通过Kubeflow pipeline 构建，学习难度高，一般与MLflow结合使用

Airflow 可视化DAG任务编排，不是mlops专用工具，可通过调用python脚本，结合MLFLOW 实现mlops





## MLflow

### Mlfow for GenAI

官方地址：https://mlflow.org/docs/latest/genai/getting-started/connect-environment/



```shell
pip install --upgrade "mlflow>=3.1" 
mlflow ui --backend-store-uri sqlite:///mlflow.db --port 5000
# 或者
mlflow ui --port 5000
```

```python
# 测试脚本
import mlflow

mlflow.set_tracking_uri("sqlite:///mlflow.db")
mlflow.set_experiment("my-genai-experiment")

# Print connection information
print(f"MLflow Tracking URI: {mlflow.get_tracking_uri()}")
print(f"Active Experiment: {mlflow.get_experiment_by_name('my-genai-experiment')}")

# Test logging
with mlflow.start_run():
    mlflow.log_param("test_param", "test_value")
    print("✓ Successfully connected to MLflow!")
```



![image-20251023125935445](https://p.ipic.vip/34llkz.png)