Search AI Products and News

Researchers from Apple and the Federal Institute of Technology in Lausanne, Switzerland, have jointly open-sourced a large-scale multimodal visual model named 4M-21. Unlike models that are optimized for specific tasks or data types, 4M-21 boasts broad versatility and flexibility. Despite having only 3 billion parameters, it offers a myriad of functionalities including image classification, object detection, semantic segmentation, instance segmentation, depth estimation, surface normal estimation

Apple and researchers from the Swiss Federal Institute of Technology (EPFL) have jointly developed a single model that can handle arbitrary-to-arbitrary modalities. This model can be trained on dozens of highly diverse modalities and perform joint training on large-scale multimodal datasets and text corpora. Named 4M-21, it is trained on 21 different modalities, completing at least three times more tasks than existing models without losing performance.