Open minmin0516 opened 1 month ago
Hello! Glad to see that you are interested in the repository.
This project is part of my Master's Thesis which is done in collaboration with a company and thus I cannot provide the data examples as that may lead to a confidentiality issue, however, I can give you a description of the data that you need to have to make the custom dataset class work.
The codes involved in the functioning of the training and testing need to be consistent when it comes to the criteria names, you can obtain that consistency by adjusting the names of the criteria in the whole stack for training and testing (You can find the names of the criteria that I use for my project but you can change it to adapt to your purpose).
This project is part of a Master's Thesis in collaboration with a company so the data used will be kept private.
The format for the train data includes:
The format of the rating headers is as follows (example with 100 images and 2 criteria):
The format of the rating values is: [values, 1, 0, 1, 1, ..., 1, 0, 1, 0, 0, 0, 0]
The format of the rating headers is: [headers, _1.png(name_of_criterion_1), _1.png(name_of_criterion_2), _2.png(name_of_criterion_1), _2.png(name_of_criterion_2), ..., _100.png(name_of_criterion_1), _100.png(name_of_criterion_2)]
Can provide training data examples, such as specific content in/path/to/labels. csv