Last week, Google quietly released an app from an AI development platform that allows users to run a wide range of AI models, which are widely available.
This app, called Google AI Edge Gallery, is available on Android and will soon be on iOS. Users can find, download and run compatible models to generate images, answer questions, write and edit code. The model runs offline without the need for an internet connection and takes advantage of supported mobile phone processors.
AI models running in the cloud are often stronger than local counterparts, but they also have downsides. Some users may want to note that they send personal or sensitive data to a remote data center or make the model available without the need to find a Wi-Fi or cellular connection.
The Google AI Edge Gallery, which Google calls the “Experimental Alpha Release,” can be downloaded from GitHub by following these instructions. The home screen displays shortcuts to AI tasks and features such as “Ask Image” and “AI Chat.” Tap the feature to see a list of models suitable for your task, such as Google’s Gemma 3N.
Google AI Edge Gallery can also provide “Prompt Lab” users that can be used to kick off “single-turn” tasks with models, such as text summaries and rewriting. The Prompt Lab comes with several task templates and configurable settings to fine-tune the behavior of your model.
Google may have different mileage in terms of performance. Modern devices with more powerful hardware will run your models faster as expected, but model size is also important. Larger models take longer to complete tasks than smaller models (for example, answering questions about images).
Google provides compelling members of the developer community with feedback on Google AI Edge Gallery Experience. This app is under the Apache 2.0 license. This means that in most contexts it can be used in commercials or other cases without limitation.
Source link
