why am i getting "WARNING:adapters.loading:Overwriting existing adapter 'adapter_name'." after training finishes?

[MRawhani]

Whenever I use AdapterTrainer object for training i get this message when training finishes:
"WARNING:adapters.loading:Overwriting existing adapter 'adapter_name'."
see this sample:

whatever task i am using it gives me this warning, why? and does it effect the performance?

[calpt]

Hi @MRawhani, this is expected if you use something like load_best_model_at_end for your training since it will re-load the best adapter checkpoint after training has finished. In that case it does not affect performance.

We'll look into making this message less confusing in the future.