Hi all! First time posting in these forums.
l’m doing a little image classification with Memes in Lesson 2, and when I show the data bunch a deal of the “Meme” text is backwards (and some of it is skewed), almost like looking at it through a mirror. See attached.
Any one have any thoughts about why that might be? More importantly, will it mess up downstream applications like classifying text?My thoughts are that I could run my model with backwards letters to train to find the backwards letters?
