Since the GPU version for the latest release still doesn’t work, I’ve been using the September GPU version. That has worked fine, even though it means giving up the Threadcount setting.
However, following a recent Windows 11 (mandatory) update, the older GPU version started failing with a 500 (internal server) error. Deepstack takes a minute to respond, and then does so with that error for every picture.
I installed the lasted CPU version (with everything else the same), and things started working normally (if slowly). Has anyone else experienced this?
Also, just to ask again, are we every going to get current CUDA versions working? I keep asking, and I keep getting no action after 7 months. I’m about ready to try another AI engine…