[Python] ML Flow 정리중

Hosting

1. 서빙

mlflow server --host 127.0.0.1 --port 9090

 

2. 추적 서버 설정

import mlflow

# Set our tracking server uri for logging
mlflow.set_tracking_uri(uri="http://127.0.0.1:9090")

# Create a new MLflow Experiment
mlflow.set_experiment("MLflow Quickstart")

# with mlflow.start_run():
	# 내부 기입되는 mlflow에 대한 로깅
    
    # Log the hyperparameters
    mlflow.log_params(params)

    # Log the loss metric
    mlflow.log_metric("accuracy", accuracy)

    # Set a tag that we can use to remind ourselves what this run was for
    mlflow.set_tag("Training Info", "Basic LR model for iris data")
    
    # ...

---

Signature

: 모델의 입/출력 & 파라미터 정보 (template) 정의

(function) def infer_signature(
    model_input: Any = None,
    model_output: Any = None,
    params: dict[str, Any] | None = None
) -> ModelSignature

 

- 예시 (python 코드 및 정의된 템플릿)

from mlflow.models import infer_signature

infer_signature(model_input={
	"long_col": 1,
	"str_col": "a",
	"bool_col": True
})

# Template (Inferred Signature)

signature:
    input: '[
        {"name": "long_col", "type": "long",    "required": "true"},
        {"name": "str_col",  "type": "string",  "required": "true"},
        {"name": "bool_col", "type": "boolean", "required": "true"}
    ]'
    output: null
    params: null

 

---