Pretrain model

[moradza]

I have two datasets, each with train, val, and test dataset. I need to pretrain model with the first dataset, and then do another round of training with the second dataset. Is there any example for this? Basically, I need to call tainer.fit(model) twice but trainer reaches max_epochs with the first dataset.

[titu1994]

You need to create a new trainer and call trainer2.fit().

[moradza]

That's what I thought. But I saw this Lightning-AI/pytorch-lightning#9636.
How can I make sure that weights are copied from the last trainer?