ASD datasets used for training DINGO. These have been accumulated through several studies and we will be adding to this data repository as we generate more ASD datasets. For now, we are not releasing ASD datasets from O4 as this contains private LVK data. There are three folders corresponding to the three different observing runs. Under each observing run folder there is a subfolder (either 8s or 16s) which sets the frequency spacing for the ASD dataset. In the O3 folder, there are both 8s and 16s ASD datasets. For now, we have not generated 16s ASD datasets for O1 and O2. The ASD datasets come in pairs labeled asds_OBSERVING_RUN_fiducial.hdf5 and asds_OBSERVING_RUN.hdf5 respectively. The _fiducial corresponds to the fact that this asd dataset only contains one ASD which used during the first stage of training. The file without the _fiducial tag indicates that ASDs sampled from across the entire observing run are used. Detailed dataset generation settings for each dataset can be found by running `dingo_ls /path/to/asd_dataset.hdf5` after installing DINGO.