Modal releases open-source training server Spindle
The story so far
Modal has released the open-source Tinker-compatible training server Spindle, which replaces token billing with GPU time billing, supports custom configurations, and hundreds of concurrent training tasks. Users can customize server configurations and pay based on GPU usage time. Modal claims that the default configuration is tuned to maximize GPU utilization and plans to continue optimizing to improve cost-effectiveness. For teams running large numbers of experiments, Spindle significantly reduces costs compared to proprietary APIs while maintaining similar performance.
AI-written from the reports · Updated 27 days ago
Coverage timeline
Follow the coverage to see every side of the story.
- Modal BlogIntroducing our Training API
Modal releases open-source Tinker compatible training server Spindle, supporting operation in Modal workspaces, replacing token billing with GPU time billing, customizable configurations, and supporting hundreds of concurrent training tasks. Spindle supports full-parameter training and multiple LoRA on various open-source models, with a distributed control plane and container auto-scaling at its core, suitable for teams needing custom training loops and cost reduction.