Pete

Thanks, makes sense. I'm still unable to get things working reliably in QuPath, including after a restart and when keeping the process alive after testing, but I'm not sure what...

Thanks @saudet good point! Alas, I'm currently stuck on macOS Big Sur 11.6.2 on (work's) M1, so can't upgrade to the latest OS - it would be nice (I think)...

It will still be possible to change the RAM, it just wouldn't be such a prominent option on startup. I'm not sure what the best default is, but I have...

Strange... if you install gradle separately, then you can also use `gradle clean jpackage` (i.e. avoid relying upon gradle wrapper). I'm not sure if that'll help. I have limited experience...

Thanks @MichaelSNelson! The main idea here is to support adding text but without really specifying what that has to be... so someone could use it in the way you describe...

Actually... the efficient way to do the clustering would perhaps be to add the number to the measurement list, and then include the 'key' in the text field of the...

I haven't tried the new classifier with a large training set, and I'm afraid I really cannot help more without a clear error message (e.g. from the log). However, the...

It won't load the pixel data if it doesn't need it, only the objects. However, it does this less efficiently than the old classifiers - and so will require *some*...

I'd need more information from you to be able to answer that... what exactly are the steps that you have applied, what features calculated, how many objects per image etc....

File format can also be significant, and whether it is read with OpenSlide/Bio-Formats (which might do some caching of their own).