Follow Datanami:
June 7, 2023

Weights & Biases Embraces the LLM Boom with Two New AI Development Tools

(Source: Weights & Biases)

Weights & Biases unveiled two new products at its inaugural in-person conference in San Francisco.

The company is known for its AI and machine learning development tools, and it previously released a new suite of LLMOps (large language model operations) tools in April to support the current LLM boom. Two new additions to this suite debuted today: W&B Weave and W&B Production Monitoring.

“W&B Weave and W&B Monitoring are the two latest additions to the product portfolio and continue to extend our support for the end-to-end ML lifecycle,” said Phil Gurbacki, VP of product at W&B. “Monitoring models is essential to helping teams maintain cutting-edge model performance in production, and ML developers need powerful, customizable tools for exploring data and building ML applications. With these two new releases, we are confident of addressing both critical areas.”

W&B Weave is an application building toolkit for data- and ML-powered applications. Weave is designed to help build AI applications through composable UI elements. There are default options for standard use cases with the ability to create custom and advanced components over time. Data scientists can also use it for creating interactive data visualizations. Weave is available as an open source LLMOps tool, and enterprise customers can find it integrated with the Weights & Biases platform.

Production Monitoring provides dashboards for visualizing AI model performance and metrics. (Source: Weights & Biases)

W&B Production Monitoring is an observability tool for ML models and ML-powered applications. The company says it enables fast troubleshooting and continuous production systems improvement. The tool logs production data and provides dashboards with analytics and graph visualizations to help users find the root causes of data drift and model degradation. There is also a custom alert feature that allows users to track custom metrics, which may appeal to LLM application developers who have unique monitoring needs. The tool also incorporates user feedback for continuous model improvement. Those interested can sign up for a private preview of Production Monitoring here.

Weights & Biases held its sold-out Fully Connected conference in San Franciso where guests heard speakers like Anyscale CEO Robert Nishihara and Richard Socher, founder and CEO of The company says its conference is dedicated to tackling the most critical topics in machine learning, including generative AI, LLMs, and production ML.

“Since our founding, we’ve been committed to building delightful, productivity-boosting tools for the ML community, and we’re thrilled to welcome thousands of ML practitioners in-person to the Fully Connected conference, our first large-scale in-person event, to explore the future of the industry,” said Shawn Lewis, CTO and co-founder, Weights & Biases. “W&B Weave and W&B Monitoring are two products that will have an immediate impact for our users and make it easier to build and monitor models over time, and we’re proud to roll these out today during the event.”

Related Items:

Weights & Biases Unveils New Suite of Tools to Support LLMOps

Anyscale Takes On LLM Challenges with Aviary

Large Language Models: Don’t Believe the Hype