I am trying to train custom OCR model. When training with my own data compared to the data provide in this tutorial, it seems to break with error message:
Local rendezvous is aborting with status: INVALID_ARGUMENT: Cannot add tensor to the batch: number of elements does not match. Shapes are: [tensor]: [4], [batch]: [3]
I tried researching, but I can not find any information regarding this situation.
Got it working. I think it was a issue of image dimensions I had to change (or perhaps some text length value)? The width and height had to match input image.