We highly recommend (re)using collection of Singularity containers prepared by ReproNim project, and ideally using datalad-container extension. See https://github.com/ReproNim/containers#a-typical-yoda-workflow for more information on how it should typically be used.
Besides using that collection via datalad, you can use them directly for the images/versions we already prefetched (if any is missing, please let us know). That might be preferable over manual fetching of container copies to not waste more space on discovery (some containers are large). E.g. to validate your BIDS dataset using bids-validator, do:
cd path/to/my/bids/dataset singularity run -B $PWD:$PWD /dartfs/rc/lab/D/DBIC/DBIC/archive/containers/images/bids/bids-validator--1.11.0.sing $PWD