I have a weighted dataset, and a model composed of two parts.
How can I train the model in such a way that the dataset weights only apply to the first part of it (while the second part is trained as if each example had the same weight)?
0 Answers
Related