Hi,
I was wondering how to create a custom learner object. I already adjusted the dataloader and coded the neural network in PyTorch (1D convolutional net). I am able to use the fit function:
fit(model,data,3,opt,F.mse_loss)
However, it would be cool to use a learner so that I can use SGDR, lr_find and all the other cool stuff:
learn.fit(lr, 3, cycle_len=1, cycle_mult=2)
I would appreciate any suggestions helping me along the way.
Thanks in advance for your help!