I think in most cases, we have a little dataset for a limited set of activates, so we don't want to change the process of feature extraction because it can make it worse due to the lack of enough data. So, it would be great to freeze the feature extraction section of the model, so their values won't be changed.
What is the feature?
Adding a field in the config file (like a Boolean), so it can be chosen to whether freeze the feature extraction part.
What is the problem this feature will solve?
I think in most cases, we have a little dataset for a limited set of activates, so we don't want to change the process of feature extraction because it can make it worse due to the lack of enough data. So, it would be great to freeze the feature extraction section of the model, so their values won't be changed.
What is the feature?
Adding a field in the config file (like a Boolean), so it can be chosen to whether freeze the feature extraction part.
What alternatives have you considered?
No response