+
Skip to content

artefom/mldag

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MLDAG

Lightweight Directed Acyclic Graphs with fit\transform support

Test status

Why useful

Most ML engineering pipelines require fit\transform methods. It would be great to be able to organize classes with fit\transform methods in graph structure and execute fit and transform separately.

MLDAG allows to do just that.

import mldag


class Preprocess:
    def fit(self, dataset):
        # Estimate some parameters
        pass

    def transform(self, dataset):
        # Apply learned transformation to dataset
        return dataset


class Model:
    def fit(self, dataset):
        # Estimate some parameters
        pass

    def transform(self, dataset):
        # Apply learned transformation to dataset
        return dataset


# Create pipeline
dag = mldag.MLDag()

# Initialize nodes
nodes = {
    'preprocess': mldag.as_node(Preprocess()),
    'model': mldag.as_node(Model())
}

dag >> nodes['preprocess'] >> nodes['model'] >> dag

dag.show()

Simple example

About

Architectural sugar for Dask out-of-memory processing workflow

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

点击 这是indexloc提供的php浏览器服务,不要输入任何密码和下载