Explorer
Content
videos
firebase-studio-now-supports-gemma-3-image-inputs-on-the-gemini-developer-api-.md
videos β€Ί firebase-studio β€Ί firebase-studio-now-supports-gemma-3-image-inputs-on-the-gemini-developer-api-.md

Firebase Studio now supports Gemma 3 image inputs on the Gemini Developer API πŸ”₯πŸͺ„πŸš€

Description

Try it out on the Gemini template and switch to the gemma-27b-it model! https://studio.firebase.google.com/new/gemini

Learn more about Gemma 3: https://ai.google.dev/gemma/docs/core

Website: https://rodydavis.com Twitter: @rodydavis Github: @rodydavis Patreon: https://www.patreon.com/rodydavis Buy Me A Coffee: https://www.buymeacoffee.com/3mgFNYd

Transcript

0:00 Β· Did you know that Firebase Studio now supports Gemma 3 with image input? You can get started by going to Firebase Studio and then going to new workspace AI and ML and get started with the Gemini API. Make sure to select JavaScript web app with V. Give a name for your workspace and hit create. Once the workspace is loaded, you can create an API key by opening up the Firebase Studio extension and then clicking on the Gemini integration.

0:29 Β· You can generate an API key and you don’t even have to leave Firebase Studio. Once it’s generated, you can copy it and paste it inside of the code.

0:38 Β· By default, the template uses Gemini 1.5 Flash. And we can test that it’s working by clicking on an image and hitting go. This will create a recipe based on the image and generate a markdown version of that recipe. But we can easily swap this out for Gemma. So go ahead to the model string and replace it with Gemma 327B instruction tuned. We can verify this is working by selecting an image and then hitting go.

1:13 Β· Try it out for yourself at studio.firebase.google.com/new/jgemini.