Hi Team, can we get the difference between L1 and L2 code difference explained in detail? I am not able to see the impact of Loss value making an impact to weights and biases.

L2-norm calculates the sum of squared differences, it penalise more when there are some outlayers in the target L1-norm: sum of ABS differences so at sterile data like in the example, where we have a uniform distribution, i reckon you dont see much difference on the result

Follow this topic