I have currently encountered an issue with the API, where when I set a queue of tasks, in my case "Upscale (2x)" with image, after an amount of jobs it suddenly changes to low-VRAM mode. Even when I set the start parameters with --always-normal-vram, it still automatically changes to low VRAM mode. As well as this, I see an unusual increase and steady use of VRAM after some jobs have completed, which probably shouldn't be happening.
And because of it, it always crashes, saying that it detected two or more devices (I currently do not have the full ERROR message).
And I don't really think it's an issue with my PC either. I have a 4070-Ti with a i9-13900K with 64GB of RAM, so PC-wise nothing should be wrong.
Additionally I'm using a pythonscript that sets a set of images for upscaling with its own prompt, settings and everything, with each generating 5 upscaled images.
I have currently encountered an issue with the API, where when I set a queue of tasks, in my case "Upscale (2x)" with image, after an amount of jobs it suddenly changes to low-VRAM mode. Even when I set the start parameters with
--always-normal-vram
, it still automatically changes to low VRAM mode. As well as this, I see an unusual increase and steady use of VRAM after some jobs have completed, which probably shouldn't be happening. And because of it, it always crashes, saying that it detected two or more devices (I currently do not have the full ERROR message). And I don't really think it's an issue with my PC either. I have a 4070-Ti with a i9-13900K with 64GB of RAM, so PC-wise nothing should be wrong.Additionally I'm using a pythonscript that sets a set of images for upscaling with its own prompt, settings and everything, with each generating 5 upscaled images.