Why isn’t this happening more already?
It takes way more resources to train the model then to use it.
It takes way more resources to train the model then to use it.