https://transformer-training-explained.vercel.app/training https://transformer-training-explained.vercel.app/training/01-training-loop https://transformer-training-explained.vercel.app/training/02-data-and-tokens https://transformer-training-explained.vercel.app/training/03-objective https://transformer-training-explained.vercel.app/training/04-backprop https://transformer-training-explained.vercel.app/training/05-optimisers https://transformer-training-explained.vercel.app/training/06-learning-rate-and-schedules https://transformer-training-explained.vercel.app/training/07-init-norm-mup https://transformer-training-explained.vercel.app/training/08-stability https://transformer-training-explained.vercel.app/training/09-mixed-precision https://transformer-training-explained.vercel.app/training/10-memory https://transformer-training-explained.vercel.app/training/11-data-parallelism https://transformer-training-explained.vercel.app/training/12-model-parallelism https://transformer-training-explained.vercel.app/training/13-slow-links https://transformer-training-explained.vercel.app/training/14-scaling-laws https://transformer-training-explained.vercel.app/training/15-running-a-real-job https://transformer-training-explained.vercel.app/training/16-after-pretraining