AI Tuning and Model Repository with Discoverability and ENS enabled use
ai.repo provides a AI fine tuning and model repository for developers, enabling discovery across fine tunes and models when public, but also providing for proprietary fine tunes and models. Developers can setup fine tunes, either as prompt based or categorization based, and then iterate on these over time using versioned sets. Those fine tunes can be shared or proprietary, and power model training. Each model is versioned as well, and can be used by others to iterate on if not set as proprietary.
This project uses polybase for login/authentication, as well as for storage of the fine tunes and models. Composedb is used for repository structure sharing and sharing/discoverability of fine tuning data and models. ENS is used to expose a ChatGPT like interface to your custom models. OpenAI is used to train models.