load-from argument in train.py

[javvi51]

It seems the load-from argument (see https://github.com/open-mmlab/mmrotate/blob/main/docs/en/get_started.md) doesn't exist (yet) in train.py, so I don't know how to fine-tune my model :( I can only resume-from a particular epoch. Any ideas or workarounds?

I guess I can increase the number of epochs in the config file, trigger early stopping and use the rest of the epochs for fine-tuning. I would need to replace the dataset in the path specified in the config file too. This is my provisional solution.