Tinker
Control every aspect of model training and fine-tuning
76 followers
Control every aspect of model training and fine-tuning
76 followers
Tinker is a flexible API for efficiently fine-tuning open source models with LoRA. It's designed for researchers and developers who want flexibility and full control of their data and algorithms without worrying about infrastructure management.
Raycast
Tinker is a flexible API for fine-tuning language models, and the first product from @miramurati's Thinking Machines lab, which also just announced their $2B seed on a $12B valuation. You'll recall that she was CTO of OpenAI, before briefly becoming CEO.
Write training loops in Python on your laptop; we'll run them on distributed GPUs. Private beta opened yesterday.
Access is limited to start, but you can sign up for the waitlist here. If you're a university or organization looking for wide scale access, contact tinker@thinkingmachines.ai.