Shortuct for traing and testing loop

[zillur-av]

Video: PyTorch workflow

I see for training and testing, a loop is required and loss function, optimizer, and everything has to be defined explicitly. I used TensorFlow before and it does all these steps in only 1-2 lines of code like model.fit(). Is there any such function in PyTorch?

[mrdbourke]

Hi @zillur-av,

Good question!

Yes there is a few options for this:

1. Create your own functions for PyTorch training/testing loops, as seen here: https://www.learnpytorch.io/05_pytorch_going_modular/
2. Use a framework such as Lightning AI or FastAI or MosaicML to do the training loop for you

Otherwise, there's no native model.fit() inside PyTorch (I'm pretty sure this is by design, to add as much hackability to PyTorch as possible).