I’m trying to follow along with a layoutlmv3 paper; however, I am using a public dataset of invoices so I could have a better understanding of how it uses layoutlmv3. At the data preparation section I keep running into a “KeyError: pixel_values” error, and whenever I try to remove the pixel_values column I get a “jpeg image file not recognized error”. The specific paper I’m following is here. The specific dataset I’m using is “mathieu1256/FATURA2-invoices” and the specific model I’m using is “microsoft/Layoutlmv3-base” because I couldn’t find the layoutlmv3-base-uncased version on hugging face and my program seems to believe it doesn’t exist.