This is more of just a weird observation than an actual problem, but thought I’d put it here just in case it is of interest (or in case the same code is reused elsewhere in cryoSPARC)
Basically, the dataset previews show what are (I think) supposed to be random images, but they very often contain duplicate particles (e.g. see attached). This seems improbable if they are being randomly selected, even if no checking for duplication is being performed. Just wanted to point it out because I found it puzzling, and wasn’t sure if it was intended.