frameleaf-ml

Find a photo by describing what you remember, and bring familiar faces together. This optional companion provides local machine learning for Frameleaf Server, including smart search and face recognition. The bundled Studio worker can also serve supported editing jobs when it shares its enrolment folder with the server. Install alongside a running Frameleaf Server on the same frameleaf Docker network. It is a background worker with no separate web interface; browse your library and manage processing from Frameleaf. Keep the model cache persistent so models do not have to download again after every restart. Model downloads need internet access and processing needs available CPU and memory. GPU use depends on a compatible image, drivers and device configuration; it is not enabled by this basic template. Mount the same Studio enrolment folder as the server, read-only here. Use Manager if you prefer guided installation and updates.

A compatible Frameleaf Server on the frameleaf Docker network; persistent model-cache storage; internet access for initial model downloads; sufficient CPU and memory for selected models. Studio worker enrolment requires the server shared folder mounted read-only. This template has no separate WebUI.