July 5, 2023

In this release, we are introducing a new dataset registration flow for improved user experience. In addition, we have added multi-media drag-and-drop support in snippets to make organization & collaboration a breeze. There are other performance improvements and fixes.

Datasets

The process for registering datasets has been completely redesigned to provide a more direct and efficient experience.

• For file uploads: Upload your file, select the delimiter, and specify the target and features to complete the process.

• For cloud reads: The number of steps has been reduced for a better UX.

Registering dataset using UI

Model Evaluation Enhancements

We have improved the model recorder to take pos_label as a parameter. This helps to identify the true positives to compute ROC/ PR curves.

from markov import EvaluationRecorder

# You can use the model_id from my_model.model_id to create
# and register EvaluationRecorder
model_id = my_model.model_id
dataset_id = markov.dataset.get_by_name("my_dataset").ds_id
evaluation_recorder = EvaluationRecorder(
    name=f"YOUR_MODEL_NAME",
    notes=f"Testing evaluation with MarkovML",
    model_id= model_id,
    # dataset_id these evaluation records belong to. 
    # You should register the evaluation dataset with MarkovML 
    # for lineage and advanced analytics. 
    dataset_id=dataset_id,
    # If you are evaluating a binary classifier, set the positive label
    # to compute the AUC/PR curve. POS_LABEL provided.
    pos_label = "POS_LABEL" 
)
# This method should be called to register a recorder with MarkovML to accept
# the incoming evaluation records.
evaluation_recorder.register()

The details are here.

Snippets

We are excited to announce a new update to Snippets. Users can now easily drag and drop images or videos onto snippets from their computers. In addition, you can link resources from various sources, such as Youtube, Loom, and Google Drive to organize all your work in one place for easy sharing and collaboration.

Performance Improvements

The embedding page has undergone significant performance improvements. With enhanced efficiency, larger datasets can now be handled with faster load and render times, making interactions with the embeddings quicker and reducing the need for prolonged loading screens.