- 博客(5)
- 收藏
- 关注
原创 An Error vs Effort ParadoxLevel 2 headings may be created by course providers in the future.
where is the mean of the output data.0 points possible (ungraded)Models that have low test error will always have values close to 1.
2023-08-12 09:49:16
78
1
原创 SGD mit
The step-sizein SGD is also known as thelearning rate. By decreasing the step sizeas the number of iterations increase, we can guarantee that SGD will converge to a local minimum. However,can't decrease too quickly or the coefficients will no longe
2023-08-10 14:53:10
42
1
空空如也
空空如也
TA创建的收藏夹 TA关注的收藏夹
TA关注的人