Google just announced the EmbeddingGemma 2 model. It is small enough to run on-device! And it is multi-modal! Meaning you can use it to search images, voice, video, in addition to just text! All secure with no API calls.
I’ll be exploring it, then teaching it on October 16th. You’ll receive all the projects discussed in the workshop, with AGENTS.md files, so you can work through each project with your own agent and apply it to your own project easily.

