Closed gekaklam closed 1 year ago
This is probably because your GUI is taking more than 10% of the VRAM on your GPU, and right now there's a hardcoded default min VRAM of 90%. You can customize this with min_gpu_mem
Maybe could update the info message to be more informative?
Hi,
I downloaded the codebase on my laptop, where I have an Nvidia GeForce RTX 3080 Mobile16GB GPU. When trying to run elicit or sweep, no matter the model that I use, pretty much always, the process is stuck with the following message:
When trying to do the same thing to the Jessica cluster, I'm getting the same, however, the message is:
In order to recreate the error, one simply needs to clone the codebase to a new env and try to run the basic commands.
Any suggestions on how to go about this?