Hi everyone, I am using some fastai V1 code:
data = ImageDataBunch.from_name_re(PATH, fnames, pat, ds_tfms=get_transforms(), size=224, bs=bs).normalize(imagenet_stats)
As I understand it, this is dividing my data (by default into 20% validation and 80% training). I would like to change this percentage and I found the
split_by_rand_pct
function.
However, I am having trouble with the syntax. I have found examples about how to use this function with ways to create the databunch but not with ImageDataBunch.from_name_re (and so far everything I have tried has resulted in different errors because I do not understand how split_by_rand_pct actually works.
Any help would be most welcome.