Function loss_fn

examples/python/logistic_regression.py:26–28

(w)

def loss_fn(w):
    logits = X @ w
    return mx.mean(mx.logaddexp(0.0, logits) - y * logits)


grad_fn = mx.grad(loss_fn)