Hi Satrio,
It is my understanding that you are working with MATLAB’s Neural Network Toolbox and are unsure about the purpose of the “revert weights” and “set weights” functions .
In the Neural Network Toolbox, “set weights” is used to manually assign specific weight and bias values to your network. This is helpful if you want to initialize the network with custom values instead of using MATLAB's default random initialization, or if you want to load weights from a previously trained model.
On the other hand, “revert weights” is used to restore the network’s weights and biases to the values they had before the most recent training attempt. This is when you want to revert back to the previously better-performing weights.
In summary, these tools give you control over the training process and help in experimenting with different weight initializations or managing model performance. For more information, refer to the following documentations:
- network: https://www.mathworks.com/help/deeplearning/ref/network.html
- revert: https://www.mathworks.com/help/deeplearning/ref/network.revert.html
- setwb: https://www.mathworks.com/help/deeplearning/ref/setwb.html
Hope this helps!
