In the docs for
ImageCleaner(), after the cleaning is done, they create a new data bunch by using
np.random.seed(42) db = (ImageList.from_df(df, path) .split_none() .label_from_df() .databunch(bs=64))
Instead of using the ‘ImageDataBunch.from_csv()’
Since the lesson doesn’t tell you how to re-start your model after you cleaned your data, I’m really lost. What is the preferred method here?