I am getting the box files in the output but they are not in the format as required by Tesseract 4.0. They are in the format of older versions of tesseract. I have checked the version which is being used for this process. It is Tesseract 4.00.00alpha. I have tried creating tiff/box file pairs using tesstrain.sh command, where I provided the text file and obtained the required tiff/box file pairs compatible to Tesseract 4.0. But this options is only for text2image. I want to create the dataset (box files) using the images i have.