What's new and changed in Watson Studio

Installing or upgrading Watson Studio

Cloud Pak for Data Version 4.7.0

A new version of Watson Studio was released in June 2023 with Cloud Pak for Data 4.7.0.

Operand version: 7.0.0

This release includes the following changes:

New features

The 7.0.0 release of Watson Studio includes the following features and updates:

Runtime 23.1 with Python and R

You can now use Runtime 23.1, which includes the latest data science frameworks on Python 3.10 and on R 4.2, to run Watson Studio Juypter notebooks, to train models, and to run Watson Machine Learning deployments.

Runtime 23.1 on R 4.2 in notebooks is supported on x86-64 hardware only.

To change environments, see Changing the environment of a notebook.

Enhanced Natural Language Processing capabilities in Runtime 23.1

Runtime 23.1 contains the new Watson Natural Language Processing library 4.1 and a new set of pre-trained models. The NLP library contains the following enhancements and updates:

• Many included models are now transformer-based. These models were trained on the Slate large language model (LLM), which was created by IBM. The models are available in two versions:
  • Optimized for CPU-only environments
  • For environments with GPUs or CPUs
• Many included models for different NLP tasks are now workflow-based instead of block-based, so you can apply the models directly on input text without worrying about preprocessing steps.
• NLP includes a Slate foundation model that you can use for fine-tuning your NLP tasks. You can use the Slate model or any transformer-based model from Hugging Face as a base to build your own models with Watson NLP.
• All models provided by IBM are now exclusively trained on unbiased data with state-of-the-art filtering for hate, bias, and profanity.

For more information, see Watson Natural Language Processing library.

New flow for adding data from a project file to a notebook

The notebook toolbar contains a new Code snippets icon that you can use to open the Code snippets pane. From the Code snippets pane, you can read data from a file or connection that was added to the project.

To generate code that inserts data to your notebook, you must now click the Code snippets icon, click Read data, and then select the data source from your project.

The Find and load data pane now serves only to upload data to a project; it does not generate any code inside the notebook.

For more information, see Loading and accessing data in a notebook.

Use JupyterLab and Juypter Notebook extensions to customize and enhance your development environment

You can now install JupyterLab and Juypter Notebook extensions to customize and enhance your development experience. Extensions can provide themes, editors, file viewers, and more. For more information, see Adding customizations to images.

Create, store, and share machine learning features

You can now speed the development of machine learning models by creating and sharing features. You add a feature group to a data asset in a project to identify the features of that data set. You can share the features with your organization by publishing the data asset to a catalog, which acts as a feature store. For more information, see Managing feature groups.

Improvements for managing your notification settings

You can now turn on Do not disturb to turn off the notifications that appear briefly in the web client.

To enable Do not disturb, click the Notifications icon () in the toolbar. Then, click the Settings icon ().

When you turn on Do not disturb, you can still see that you have unread notifications on the Notifications icon () in the toolbar.

For more information, see Setting your notification preferences.

Use connections from different Cloud Pak for Data instances in Git-based projects

Git-based projects can be imported to multiple instances of Cloud Pak for Data. To ensure that you can access the same data from different instances of Cloud Pak for Data, you can create connections that are based on a copy of a platform connection. A connection that is based on a copy of a platform connection can be used across instances of Cloud Pak for Data. For more information, see Connecting to data sources in a Git-based project.

Removal of Scala environments

All runtime environments based on the Scala programming language have been removed.

Security fixes

This release includes fixes for the following security issues:

CVE-2023-31125, CVE-2023-27535, CVE-2023-23931, CVE-2023-0361, CVE-2023-0286

CVE-2022-46175, CVE-2022-45688, CVE-2022-43552, CVE-2022-35252

CVE-2021-44906

Customer-reported issues fixed in this release

DT214953: JSpark UI not working – 502 bad gateway

DT212584: Setting HIDE_CATALOG_CONNECTIONS_FROM_ASSET_BROWSER to true in the product-configmap has no effect in cp4d 4.6.3

Bug fixes

This release includes the following fixes:

Incorrect password imports project successfully with falsely decrypted properties and you don't receive an error message

• Issue: If the exported file that you select to import was encrypted, you must enter the password that was used for encryption to enable decrypting sensitive connection properties. If you enter the incorrect password to import a local file, an error message is not received and the file imports successfully with falsely decrypted sensitive connection properties.
• Resolution: You will receive a warning if the password you provide does not match the password that was used for exporting the archive. Encrypted data will not be properly decrypted with the incorrect password.

Downloading a data asset from a Cloud Object Storage connection can result in a timeout

• Issue: Downloading a data asset from a Cloud Object Storage connection will timeout if the source connection was created without specifying a bucket, secret key, and access key.
• Resolution: You receive an error indicating that a bucket or HMAC credentials are needed to proceed, and the download no longer times out.