Prepared the Optimization documentation covering SGD, momentum, Adam, AdamW, learning-rate scheduling, weight decay, gradient clipping, gradient accumulation, optimizer state, convergence, numerical stability, performance, and testing.