Difference between revisions of "Early Stopping"
m |
m |
||
| Line 13: | Line 13: | ||
* add more data | * add more data | ||
| − | * use [[Data | + | * use [[Data Quality#Batch Norm(alization) & Standardization|Batch Norm(alization) & Standardization]] |
| − | |||
* use architectures that generalize well | * use architectures that generalize well | ||
* reduce architecture complexity | * reduce architecture complexity | ||
* add [[Regularization]] | * add [[Regularization]] | ||
** [[L1 and L2 Regularization]] - update the general cost function by adding another term known as the regularization term. | ** [[L1 and L2 Regularization]] - update the general cost function by adding another term known as the regularization term. | ||
| − | ** | + | ** Dropout - at every iteration, it randomly selects some nodes and temporarily removes the nodes (along with all of their incoming and outgoing connections) |
| − | ** [[Data Augmentation]] | + | ** [[Data Augmentation, Data Labeling, and Auto-Tagging|Data Augmentation]] |
| − | ** Early Stopping | + | ** [[Early Stopping]] |
<youtube>7QfUNxkthq8</youtube> | <youtube>7QfUNxkthq8</youtube> | ||
<youtube>ATuyK_HWZgc</youtube> | <youtube>ATuyK_HWZgc</youtube> | ||
Revision as of 19:19, 19 September 2020
Youtube search... ...Google search
Good practices for addressing the Overfitting Challenge:
- add more data
- use Batch Norm(alization) & Standardization
- use architectures that generalize well
- reduce architecture complexity
- add Regularization
- L1 and L2 Regularization - update the general cost function by adding another term known as the regularization term.
- Dropout - at every iteration, it randomly selects some nodes and temporarily removes the nodes (along with all of their incoming and outgoing connections)
- Data Augmentation
- Early Stopping
