Weights & Biases Announces Major Updates for Its MLOps Platform 

SAN FRANCISCO, Oct. 18, 2022 -- Weights & Biases today announced significant enhancements to its developer-first Machine Learning Operations (MLOps) platform. The latest additions will accelerate ML activities for enterprises and ML practitioners, providing a seamless, end-to-end MLOps experience from model prototyping to production.

There seems to be a new large ML model grabbing headlines every week. Whether it’s OpenAI’s big releases like GPT-3, Dalle-2, or Whisper, or one of the many open-source projects generating state-of-the-art models, like Stable Diffusion, OpenFold, or Craiyon, these models have found their way into the mainstream. All of the ML teams developing these models use Weights & Biases to debug their models, collaboratively improve them, and accelerate their work.

“As enterprises invest heavily in machine learning infrastructure and workflows become more complex, today’s ML practitioners need a platform that will work across the entire ML lifecycle,” said Shawn Lewis, CTO and Co-Founder, Weights & Biases. “We’re committed to making machine learning work in the real world, all within a platform that will increase productivity and output. By expanding use case coverage, offering new deployment options, adding ML workload orchestration, increasing security investment, and much more, the latest enhancements build on our mission to help all customers succeed with ML.”

In the latest version of the platform, Weights & Biases unveiled:

• W&B Models – W&B Models, now available on public preview, is a scalable way to govern ML model lifecycles in a central repository and enable cross-functional discovery and collaboration. With reproducibility and lineage tracking capabilities that allow users to track exactly when a model moved from staging to production, W&B Models ensures customers maintain and ensure a high quality of models in production over time.
• W&B Launch – The ultimate tool for ML workload orchestration, W&B Launch allows users to build reproducible ML workflows for agility, team efficiency, and optimized resource utilization. Containerized workflow (Docker and Kubernetes) provides easy reproducibility in any environment, at scale, and rule-based, time-based triggers CI/CD automation of ML workflow.
• Time Series Support – New Time Series Forecasting will allow users to optimize the accuracy of complex time series models, including the ability to quickly back-test performance, compare with baseline model performance, and visualize the results in the format and manner that users are accustomed to.
• Enhanced 3D Bounding Boxes – W&B continues to expand its first-class support of media types by enhancing 3D Object and Point Cloud logging. Users can log 3D objects and light detection and radar (LiDAR) point clouds, including bounding box annotations, prediction scores, and camera viewpoint.  Autonomous vehicle companies, for example, can dynamically visualize LiDAR point clouds, annotate with 3D bounding boxes, and explore interactively with Weights & Biases.
• W&B Profile Pages – New W&B Profile Pages let users and teams customize their profile page and build an “ML portfolio” to easily share their work, show their skills, and demonstrate their credibility with the world. Like a GitHub profile for code, the new profiles allow users to control which information is publicly available and customize which projects are highlighted.
• Dedicated Cloud – For customers with sensitive use cases or stringent enterprise security requirements, W&B’s Dedicated Cloud provides a secure and flexible managed environment. The company has added support for Microsoft Azure, and now supports managed installations for the three major cloud providers – Azure, GCP, and AWS – to power Enterprise ML Lifecycle Management. Weights & Biases is also available on the AWS, Azure, and GCP Marketplaces, making it even easier for customers to purchase the platform.
• Enterprise Security Portal – The new security portal offers a centralized hub to access all W&B security frameworks, regulations, and certifications. This ensures fast compliance with the most stringent security standards – integral for large customers in highly regulated industries.

Corporate Milestones

The platform enhancements build on tremendous traction that Weights & Biases achieved in the last 12 months. The company grew its employee base by more than 80% year over year and expanded operations to EMEA, including adding go-to-market teams in the UK and Germany. On the adoption front, Weights & Biases’ Annual Recurring Revenue (ARR) grew approximately 3x year over year and its net promoter score topped 75.

“We’re dedicated to building and innovating the best products for ML – everything we do is focused on unlocking productivity by optimizing, visualizing, and standardizing model and data pipelines, regardless of framework, environment, or workflow,” said Phil Gurbacki, VP of Product at Weights & Biases. “With strong adoption and satisfaction from our customers to date, we look forward to continuing to delight our users. From more flexible deployment options to customized profiles and the ability to orchestrate ML workloads, today’s new introductions are a significant step in our journey to becoming the system of record for all ML practitioners.”

To learn more about the Weights & Biases developer-first MLOps platform and additional enhancements not highlighted, visit wandb.ai.

About Weights & Biases

Weights & Biases is the leading developer-first MLOps platform that provides enterprise-grade, end-to-end MLOps workflow to accelerate ML activities. Used by top ML practitioners including teams at NVIDIA, OpenAI, Lyft, Blue River Technology, Toyota, and MILA, Weights & Biases is part of the new standard of best practices for machine learning.

Source: Weights & Biases