Everybody calls it. Far fewer can say what the bias correction is for, which is a shame, because it is the reason your first hundred steps do not go nowhere.
Write adam_step(params, grads, m, v, t, lr=1e-3, beta1=0.9, beta2=0.999, eps=1e-8) returning (new_params, new_m, new_v).
m = beta1*m + (1-beta1)*g
v = beta2*v + (1-beta2)*g²
m̂ = m / (1 - beta1^t)
v̂ = v / (1 - beta2^t)
params = params - lr * m̂ / (sqrt(v̂) + eps)
t is the step number and starts at 1. Return the updated moments too: the caller passes them back next step.
Build the architecture on a canvas: place the components, configure them, connect them into a data flow, and write a short reason for each one. The AI reviewer grades your design against a rubric written specifically for this problem.
Minimum 5 components · needs a wide desktop screen