This is a specialised version of the official pretrained TiRex-2 model which we trained on datasets that do not overlap with Gift-Eval but the training included Gift-Eval-Pretrain. For more information please see TiRex-2.
