I have set up a face detection by gender workflow as shown in the picture. If I run batches of 2 or more, I noted that on occasion some faces from other images would be pasted onto the wrong image. It's no problem to just run with one latent at a time, so no worries, just wanted to flag the issue.
