PROTRAILBLAZER

Gradient Descent · Loss Landscape
step0
loss0.000
|grad|0.000
x,y0, 0
current position
global minimum
path history
The optimizer follows the negative gradient of the loss surface, stepping downhill toward a minimum. Adjust the learning rate to see how step size affects convergence.
Learning Rate
0.035