Deep Learning Concept

Loss Functions · MSE vs Cross-Entropy

Drag the prediction toward — or away from — the true label, and compare how each loss reacts.

prediction p loss y
MSECross-Entropy
prediction p
0.70
MSE = (p−y)²
0.090
Cross-Entropy
0.357
true label y=1 · move p toward 0 and watch CE explode
loss.py