Mira Murati’s AI startup, Thinking Machines Lab, has unveiled Tinker, a flexible Application Programming Interface (API) designed for fine-tuning language models (LLMs). Targeting researchers and developers, Tinker simplifies experimentation by providing control over algorithms and data while managing the complexities of distributed training. “Tinker advances our mission of enabling more people to customize cutting-edge models,” Murati stated on X. This managed service allows seamless transitions between lightweight models and advanced systems with minimal code changes. Currently in private beta, Tinker offers free initial access, with usage-based pricing coming soon. Notable research teams from Princeton, Stanford, Berkeley, and others have tested the platform, exploring various AI applications. To further assist developers, Thinking Machines Lab is launching the Tinker Cookbook, an open-source library of modern post-training methods built on the Tinker API. Explore how Tinker could revolutionize AI model fine-tuning and enhance your research today!
Source link
