7/ ConvNets Match Vision Transformers – evaluates a performant ConvNet architecture pretrained on JFT-4B at scale; observes a log-log scaling law between the held out loss and compute budget.

→ View original post on X — @dair_ai