Hugging Face for everyone How you can try out Apple's AI model 4M for yourself
Apple is now taking a further step and allowing its users to participate directly. But what can the model do and what can't it do? A look at Hugging Face.
No time? blue News summarizes for you
- Apple has published a demo version of its AI model 4M on the Hugging Face platform.
- Among other things, the 4M model can generate images from text, process 3D scenes using natural language and recognize complex objects in images.
- Users can test the demo version directly on Hugging Face by uploading or recording images.
Apple has released a demo version of its new AI model "4M", which is now available on the Hugging Face platform.
According to the description on Hugging Face, the model is not only scalable and open source, but also works across different modalities and tasks. Put simply, the model is able to process and manipulate data from different disciplines. For example, it can generate images from text, process 3D scenes using natural language or recognize complex objects in images.
How you can test it yourself
Interested parties can try out the demo version directly on Hugging Face. Users have the option of uploading an image, recording it with the webcam or pasting it from the clipboard.
By clicking on the "Predict with 4M-21 XL" button, the model attempts to recognize all modalities of the image. This includes the RGB layer, depth information, the color palette and objects such as people, windows and walls. It also creates a caption for the image, recognizes metadata and other layers.
The demo impressively shows that the AI is able to recognize multimodal layers in images. In practice, the model could be used to add new functions to Apple's software. These include, for example, the processing of images and videos using natural language. The demo version thus provides a promising outlook on the possibilities and potential of Apple's new AI model.
