Last week, Google discreetly published an application that allows users to run a range of AI models openly available from the face of the AI Dev platform in their phones.
Called Google Ai Edge galleryThe application is available for Android and will soon arrive on iOS. It allows users to find, download and run compatible models that generate images, answer questions, write and modify the code, etc. The models are running out of line, without the need for an internet connection, explaining the supported phones processors.
AI models operating in the cloud are often more powerful than their local counterparts, but they also have their drawbacks. Some users may be wary of sending personal or sensitive data to a remote data center, or wish to have models available without the need to find a Wi-Fi or cellular connection.
Google Ai Edge Gallery, whom Google calls an “experimental alpha version”, can be downloaded from Github by following These instructions. The home screen displays shortcuts to AI tasks and capacities as “ask for the image” and “Chat IA”. The tapping on a capacity draws a list of models adapted to the task, such as GEMMA 3N of Google.
Google Ai Edge Gallery also provides a “fast laboratory” that users can use to launch the “fire tower” tasks powered by models, such as the summary and the rewriting of the text. The promoting laboratory is delivered with several models of tasks and configurable parameters to refine the behavior of the models.
Your mileage can vary in terms of performance, warns Google. Modern devices with more powerful equipment will predict models faster, but the size of the model is also important. The larger models will take more time to accomplish a task – for example, by answering a question about an image – than smaller models.
Google’s inviting members of the developer community to give comments on the Google AI Edge Gallery experience. The application is under a Apache 2.0 license, which means that it can be used in most of the contexts – commercial or not – without restriction.
