How to resume from checkpoint without load optimiser

[yikuanli]

I have a pre-training model, while I just want to load the model and train it with a different optimiser, I don't know how to resume from the checkpoint without load the optimiser using pytorch_lightning

[hiyyg]

I think you need to manually remove the solver and scheduler states, like https://github.com/aim-uofa/AdelaiDet/blob/master/tools/remove_optim_from_ckpt.py.