https://preview.redd.it/3ccm5gd1puzg1.png?width=1179&format=png&auto=webp&s=c940d2e6ef1d61288ac214eae4679a7c910b7917 Today, I’m talking about a new research paper from Token AI: "Stable Training with Adaptive Momentum" It introduces what could be one of the strongest optimizers, both in theory and i
