Data Science Lab
- Auto ML Integration:
- There will be an option to train models with different algorithms and get the best model among them using Auto ML applications.
- A new tab along with datasets is provided at the project level for Auto ML experimentations.
- Data Preparation Integration:
- Integrated the Data Preparation services in the DS Lab module.
- Provided an icon in the Datasets tab to generate Data Preparations.
- Pytorch inferencing in Pipeline:
- A Pytorch model can be trained, saved, and registered inside the Data Pipeline module.
- The registered model can infer with the test data and get the predictions in the DS Lab model runner.
- Data Catalog Implementation:
- Displays the consolidated details of Data science projects, Notebooks, models, and Utility scripts.
- Provided illustrative representation in UI.
- Kernel Session Manager: Introduced a Kernel Restart icon on the Notebook header panel to kill the current session and create a new one.
- The Output cell height has been increased to 500px.
- Ellipsis Icon for Cell Operations: Provided the cell operations inside the Ellipsis icon to enhance the UI experience.
- Timestamp for Notebook and Model lists: Provide Timestamp as a separate entity in the list to identify the last updated time from UI.
- Algorithms:
- To remove the extra Print statements and unwanted output parameters.
- Introduced optimized and standardized code.
- Merge and Split cells: Two or more cells get split or merged inside the DS Lab Notebook page.
- Static Update of the List: Instead of refreshing the entire list, a loader appears at the entity level to refresh it.
- Save As Operation: The Save As option has been provided in the Notebook toolbar to create a copy of the DS Notebook.
- Artifacts enhancement: Provided a Preview option in the Artifacts UI with limited rows. In case of the images, the .png file itself can be previewed.
- Data Sets Enhancement: Provided a Preview option in the UI with limited rows to preview the added data set.
- Provided an Add option inside the Notebook to directly add Data Sets and Data Sandbox files.
- Variable Explorer: Provided a variable explorer to show the name, type, length, and sample values for all variables that are created in the Notebook.
Please Note: The Auto ML is an alpha release.