Vivid

Support Tools for Machine Learning Vividly 🚀

Documents

Usage

The concept of vivid is easy to use. Only make instance and run fit, vivid save model metrics and weights (like feature_imporance, pr/auc curve, training time, ...) .

import pandas as pd
from sklearn.datasets import load_boston

from vivid.backends.experiments import LocalExperimentBackend
from vivid.estimators.boosting import XGBRegressorBlock

X, y = load_boston(return_X_y=True)
train_df = pd.DataFrame(X)

# create model and experiment
xgb = XGBRegressorBlock('xgb')
experiment = LocalExperimentBackend('./outputs/simple')

# run models
from vivid.runner import create_runner

runner = create_runner(blocks=xgb, experiment=experiment)
runner.fit(train_df, y)
runner.predict(train_df)

VIVID makes it easy to describe model/feature relationships. For example, you can easily describe stacking, which can be quite complicated if you create it normally.

Install

pip install python-vivid

Sample Code

In /vivid/samples, Some sample script codes exist.

Developer

Requirements

• docker
• docker-compose

create docker-image from docker-compose file

docker-compose build
docker-compose up -d
docker exec -it vivid-test bash

Test

use pytest for test tool (see gitlab-ci.yml).

pytest tests