Closed anlava closed 1 year ago
DVC pipeline for training and validation of three models: Random forest, CatBoost and MLP.
DVC pipeline structure (dvc dag command output):
dvc dag
+----------------------------------------------------+ +-------------------------------------+ | data/light-curves/generated_1m_cleaned.parquet.dvc | | data/light-curves/lc_1M.parquet.dvc | +----------------------------------------------------+ +-------------------------------------+ **** **** ***** ***** *** *** +------------------+ ******| prepare_datasets |******* ********************+------------------+ ****************** ************** ******* ** ** ****** ************ ************** ******* *** ** ******* ************ ******* ******** ** ** **** ************ +----------+ ***** +-----------+ ** +----------------+ ******* | train_rf | ** | train_mlp | ** | train_catboost | ** +----------+ ** +-----------+ ** +----------------+ ** ** ** ** ** *** *** ** ** ** ** ** ** ** ** ** ** ** ** +-------------+ +--------------+ +-------------------+ | evaluate_rf | | evaluate_mlp | | evaluate_catboost | +-------------+ +--------------+ +-------------------+
To start pipeline run dvc repro command. All parameters contained inparams.yaml file
dvc repro
params.yaml
Looks great! Could you please add a short instruction on how to run it to the Readme?
DVC pipeline for training and validation of three models: Random forest, CatBoost and MLP.
DVC pipeline structure (
dvc dag
command output):To start pipeline run
dvc repro
command. All parameters contained inparams.yaml
file