Introducing W&B Serverless SFT: Fine-tune LLMs with no infra headaches

Core Components

• Experiments
  Track and visualize your ML experiments
• Sweeps
  Optimize your hyperparameters
• Tables
  Visualize and explore your ML data
• Reports
  Visualize and explore your ML data

Training Options

• Serverless RL
  Fine-tune LLMs without managing GPUs
• ART
  Open-source RL framework
• Ruler
  Automated reward function for RL

Inference Models

• OpenAI OSS
  GPT OSS 20B, GPT OSS 120B
• Alibaba Qwen3
  23B A22B, 23B5B Thinking, Coder 480B
• Meta Llama
  Llama 4 Scout, 3.3 70B, 3.1 8B
• MoonshotAI Kimi
  Kimi K2
• Microsoft Phi
  Phi 4 Mini 3.8B
• Hangzhou DeepSeek
  DeepSeek V3.1, V3-0324, R1-0528
• Z.ai
  Z.AI GLM 4.5

Weave Tools

• Traces
  Explore and debug AI applications
• Evaluations
  Rigorous evaluations of AI applications
• Playground
  Explore prompts and models
• Agents
  Observability tools for agentic systems
• Guardrails
  Block prompt attacks and harmful outputs
• Monitors
  Continuously improve in production

Use Cases

• Train LLMs
• Fine-tune LLMs
• Computer Vision
• Time Series
• Recommender Systems
• Classification & Regression

Industries Served

• Autonomous Vehicles
• Communications
• Financial Services
• Healthcare & Life Sciences
• Public Sector
• Scientific Research

Case Studies

• Canva
  Learn how Canva leverages W&B to deploy models
• Microsoft
  Learn how Microsoft uses W&B for their ML projects
• Toyota
  Learn how Toyota uses W&B for autonomous driving
• OpenAI
  Learn how OpenAI Robotics uses W&B for large scale ML

Additional Resources

• AI Courses
• Blog
• Articles
• Podcast
• Whitepapers
• Events & Webinars
• Press

Learn more about Weights & Biases and their offerings to enhance your ML projects.