Early stopping feature

[nupoor-ka]

Search before asking

• I have searched the Multimodal Maestro issues and found no similar feature requests.

Question

Hi, I've been using Maestro to finetune Florence2 on a custom dataset, it's a great tool. One issue I've faced is that I have to guess at the number of epochs I should run fine-tuning for instead of allowing for an early-stop. Is there any way to set it up to stop at a certain minimum loss value or stop when the change in training loss reaches a certain limit?

Additional

No response