Christian Bartz
Christian Bartz
An ordered dictionary is the expected output =) The keys of this dictionary are the predicted words and the values are ordered dictionaries with the bboxes predicted for this word.
You are using an image, where the network does not know how to handle it. It has only been trained on images with one line of cropped-text, and not two....
Well, you did not separate the digits in the first line by a tab (`\t`). The `csv` needs to be a `tsv`.
okay, did you also change the delimiter between all values in the `tsv` file? All items need to be delimited by a tab character.
Hi, I'm sorry but I can not reproduce you problems. Could you give me some more information?
Okay, that is very interesting. I can access all files, download them and extract them without any problem. I just tried right now. Please try again. If it does not...
Hmm, you are right, some images only contain 0 bytes. Maybe you should open an issue in that [repository](https://github.com/ocr-algorithm-and-data/SVT). If they do not help you, you'll have to download and...
No, there is currently no code available that supports this, but you can create a script for that if you like :wink:
Yes, you will need the char_map in order to map the predicted classes back to characters.
You can find the `char_map` [here](https://github.com/Bartzi/kiss/blob/master/train_utils/char-map-bos.json) :smiley: