Check.

Getting DS4 to run at a reasonable speed was pretty tricky, GLM 5.3 a lot trickier because if you don't want to have a model that is quantized too far down that is a fortune in VRAM and GPUs at today's prices.