Integrations
Deepchecks
and
ZenML logo in purple, representing machine learning pipelines and MLOps framework.
Automate robust data and model validation in your ML pipelines with Deepchecks and ZenML
Deepchecks
All integrations

Deepchecks

Automate robust data and model validation in your ML pipelines with Deepchecks and ZenML
Add to ZenML
COMPARE
related resources
No items found.

Automate robust data and model validation in your ML pipelines with Deepchecks and ZenML

The Deepchecks integration enables you to seamlessly incorporate comprehensive data integrity, data drift, model drift and model performance checks into your ZenML pipelines. By leveraging Deepchecks' extensive library of pre-configured tests, you can easily validate the quality and reliability of the datasets and models used in your ML workflows, ensuring more robust and production-ready pipelines.

Features with ZenML

  • Seamlessly integrate Deepchecks tests into ZenML pipelines using pre-built steps
  • Automatically validate data integrity, detect data drift, evaluate models and compare model performance
  • Visualize interactive test results and reports directly in ZenML artifacts and dashboard
  • Implement test result based branching and error handling for more robust pipelines
  • Switch between different levels of integration to maximize flexibility

Main Features

  • Extensive library of pre-configured data validation and model evaluation tests
  • Supports both tabular data and computer vision use cases
  • Smart defaults allow running test suites with minimal configuration
  • Fully customizable test conditions and validation logic
  • Generates interactive visual reports for easier analysis and sharing

How to use ZenML with
Deepchecks

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.base import ClassifierMixin
from sklearn.ensemble import RandomForestClassifier

from zenml import pipeline, step
from zenml.integrations.constants import DEEPCHECKS, SKLEARN

from deepchecks.tabular.datasets.classification import iris
from typing_extensions import Tuple, Annotated

from zenml.artifacts.artifact_config import ArtifactConfig

LABEL_COL = "target"

@step
def data_loader() -> Tuple[
    Annotated[
        pd.DataFrame, ArtifactConfig(name="reference_dataset")
    ],
    Annotated[
        pd.DataFrame,
        ArtifactConfig(name="comparison_dataset"),
    ],
]:
    """Load the iris dataset."""
    iris_df = iris.load_data(data_format="Dataframe", as_train_test=False)
    df_train, df_test = train_test_split(
        iris_df, stratify=iris_df[LABEL_COL], random_state=0
    )
    return df_train, df_test

@step
def trainer(df_train: pd.DataFrame) -> Annotated[ClassifierMixin, ArtifactConfig(name="model")]:
    # Train Model
    rf_clf = RandomForestClassifier(random_state=0)
    rf_clf.fit(df_train.drop(LABEL_COL, axis=1), df_train[LABEL_COL])
    return rf_clf

from zenml.integrations.deepchecks.steps import (
    deepchecks_data_integrity_check_step,
)

data_validator = deepchecks_data_integrity_check_step.with_options(
    parameters=dict(
        dataset_kwargs=dict(label=LABEL_COL, cat_features=[]),
    ),
)

from zenml.integrations.deepchecks.steps import (
    deepchecks_data_drift_check_step,
)

data_drift_detector = deepchecks_data_drift_check_step.with_options(
    parameters=dict(dataset_kwargs=dict(label=LABEL_COL, cat_features=[]))
)

from zenml.integrations.deepchecks.steps import (
    deepchecks_model_validation_check_step,
)

model_validator = deepchecks_model_validation_check_step.with_options(
    parameters=dict(
        dataset_kwargs=dict(label=LABEL_COL, cat_features=[]),
    ),
)

from zenml.integrations.deepchecks.steps import (
    deepchecks_model_drift_check_step,
)

model_drift_detector = deepchecks_model_drift_check_step.with_options(
    parameters=dict(
        dataset_kwargs=dict(label=LABEL_COL, cat_features=[]),
    ),
)

from zenml.config import DockerSettings
docker_settings = DockerSettings(required_integrations=[DEEPCHECKS, SKLEARN])

@pipeline(enable_cache=True, settings={"docker": docker_settings})
def data_validation_pipeline():
    """Links all the steps together in a pipeline"""
    df_train, df_test = data_loader()
    data_validator(dataset=df_train)
    data_drift_detector(
        reference_dataset=df_train,
        target_dataset=df_test,
    )
    model = trainer(df_train)
    model_validator(dataset=df_train, model=model)
    model_drift_detector(
        reference_dataset=df_train, target_dataset=df_test, model=model
    )


if __name__ == "__main__":
    # Run the pipeline
    data_validation_pipeline()
    
    

In the code above, Deepchecks is integrated with ZenML to perform various data validation and model validation checks. The data_loader step loads the Iris dataset and splits it into training and testing sets. The trainer step trains a RandomForestClassifier on the training data. The deepchecks_data_integrity_check_step is used to validate the integrity of the training data, while the deepchecks_data_drift_check_step detects any data drift between the training and testing datasets. The deepchecks_model_validation_check_step validates the trained model on the training data, and the deepchecks_model_drift_check_step checks for model drift between the training and testing datasets. These steps are linked together in a ZenML pipeline, which is configured to use Docker settings for the required integrations. This setup ensures comprehensive data and model validation using Deepchecks within a ZenML pipeline.

Additional Resources
Example: Data Validation Pipeline with Deepchecks
Deepchecks Integration Docs
Deepchecks Website

Automate robust data and model validation in your ML pipelines with Deepchecks and ZenML

The Deepchecks integration enables you to seamlessly incorporate comprehensive data integrity, data drift, model drift and model performance checks into your ZenML pipelines. By leveraging Deepchecks' extensive library of pre-configured tests, you can easily validate the quality and reliability of the datasets and models used in your ML workflows, ensuring more robust and production-ready pipelines.
Deepchecks

Start Your Free Trial Now

No new paradigms - Bring your own tools and infrastructure
No data leaves your servers, we only track metadata
Free trial included - no strings attached, cancel anytime
Dashboard displaying machine learning models, including versions, authors, and tags. Relevant to model monitoring and ML pipelines.

Connect Your ML Pipelines to a World of Tools

Expand your ML pipelines with Apache Airflow and other 50+ ZenML Integrations
Elastic Container Registry
Comet
Weights & Biases
PyTorch
Databricks
BentoML
Slack
MLflow
Amazon S3
Google Cloud Storage (GCS)
Argilla