Hi!
From the lecture, we run gradient descent until convergence. What does convergence mean? For example, in the practice lab in the code cell 12, which line of code indicates about convergence?
Jus want to understand the concept more. Thank you.
Convergence means that we’ve found the minimum cost.
I think I found my answer from a slide. Let epsilon be 0.001. If J_wb (cost) decreases by <= epsion in one iteration, declare convergence. Thanks.
1 Like