Batch Gradient Descent: Difference between revisions
From Rice Wiki
No edit summary |
No edit summary |
||
| Line 4: | Line 4: | ||
* Less performing/precise (not always) | * Less performing/precise (not always) | ||
A variation, '''mini batch GD,''' uses smaller batches (not the entire dataset). It mitigates the lack in precision. | A variation, '''mini batch GD,''' uses smaller batches (not the entire dataset). It mitigates the lack in precision. | ||
[[Category:Machine Learning]] | |||
Latest revision as of 19:31, 17 May 2024
In batch gradient descent, the unit of data is the entire dataset, in contrast to Stochastic Gradient Descent whose unit of data is one data point. It uses the average of the computed gradients to update the weights of a batch of data points.
- Faster
- Less performing/precise (not always)
A variation, mini batch GD, uses smaller batches (not the entire dataset). It mitigates the lack in precision.
