Hacker News new | past | comments | ask | show | jobs | submit
smallerize | on: Making deep learning go brrrr from first principles (2022)
Does this mean that if your model is "overfitting", the solution is to train for even more epochs?
loading story #48252096