March 4 Product Update: Regression Baseline Models, Custom Metadata, and Embedding Controls

Our latest release comes packed with several improvements that are aimed at making your experience smoother.

Support for Regression Baseline Models

We're excited to share that we now support Regression Baseline Models. MarkovML will automatically create a Regression Baseline model for datasets with continuous targets. This feature will empower you to establish a foundational benchmark for regression tasks, paving the way for more advanced model development.

Here is an example of an app that uses a Money Laundering dataset, where the users can input fields like credictScore and taxationScore, and receive the kycRiskScore as an output.

SDK Enhancements

🚧 Update SDK version to v2.0.4

Please update SDK to latest version. Refer Upgrade MarkovML SDK .

Note: SDK versions v1.x.x and older are no longer supported. Please make sure you are on the latest version.

Add Custom Metadata to Model in Markov SDK

Markov SDK now supports adding custom metadata to models, providing enhanced flexibility in managing your models. This update will allow you to add custom metadata to themodels and retrieve models using metadata filters as shown below:

Python

# Add custom metadata to your  model
existing_model.add_metadata(metadata={"tag": "latest", "comment": "good"})

# Retrieve your model using metadata filters
model_by_metadata = Model.get_models_by_metadata(metadata={"tag": "latest"})
    

Refer to Building your ML Models page to learn how to build an ML model using Markov SDK.

Improved Embeddings Controls

We've streamlined the controls on the embedding screen for a more user-friendly experience. The extensive controls have been replaced with a simple gear icon ⚙️, consolidating all settings related to custom embeddings in one accessible location.

Bug Fixes and Performance Improvements

Additionally, we have made several bug fixes and performance improvements to make your experience with Markov better than ever before.