# Multimodal adaptation

A 7-billion-parameter vision Transformer model trained on 8 billion language-unlabeled web images, achieving exceptional visual representation capabilities through self-supervised learning

This is a tiny model for debugging, randomly initialized based on the adjusted configuration, specifically designed for rapid process verification.

AIM-v2 is an image encoder model based on the timm library, with a scale of 1 billion parameters, suitable for image feature extraction tasks.