Any similar function?

Is there any function in Pytorch similar to history in Tensorflow?

What does this method do?
I could only find a Keras History function, which seems to be some kind of callback.

Yes I am pretty sure thats what they meant. It records loss, lr, score, etc. So you have everything you need at the end of a fold for example to make a nice graph. Pytorch doesn’t have anything like Keras history does it?

PyTorch doesn’t ship with these utilities (the ImageNet example uses some metric objects, but these are probably not what you are looking for).
However, higher-level wrappers such as Ignite or Lightning provide such callbacks.