Increases data size

In lesson two, when I train the model on smaller images and get ready to train them on larger I pass: learn.set_data(get_data(299,bs))

but if I directly pass, 3,cycle_len=1) after this I get an error. I have to pass learn.freeze() before fitting. What is the reason for this?

Sounds like you might be running out of memory on your GPU. The error might contain some useful information for troubleshooting this.

I am using crestle for GPU