Is it allowed to use open source softwares or pretrained models?
For example, is it allowed to use tesseract in the txtocr problem?
The rule description says
Any use of an external public dataset is not allowed; participants must use the datasets made available by the AIcrowd for the purpose of training and testing their model.