Imbalanced-data
The dataset has 4,795 classes, with a significant imbalance between the numbers of samples in these classes, which ranges from one sample to several hundred samples. The training and test splits include segmented samples extracted from 1000 documents.
You can download the dataset and pretrained weights from the following link address. https://drive.google.com/file/d/1DxMQFCeZHWKCws5xrPXt6pt09rvGWof3/view?usp=sharing https://drive.google.com/file/d/1__hj1hL3mSRwflEgff3mYmfj54Et8_vx/view?usp=sharing
There are two folders, train and test. The subfolders in each folder represent the classes.