SethRobinson
feature request: progress status & interrrupt, free VRAM
Hi Seth,
may I ask you some other questions?
I've searched the scripts but could not find a option to get the current progress (single and total progess) information. I mean the information which is shown in the server window with the following details: "percentage (0-100), steps (0/20), time, estimated time.
What I also could not find, is something like the interrupt function. In the web ui there is a button called "interrupt" (it appears while generating txt2img). Is there by chance a way to stop / abort / restart the generation on the server and free GPU VRAM? From time to time the server returns this error: "...CUDA out of memory, ...! The only way to fix this, is to restart the server? Is there a option to free VRAM (memory allocation) from inside ai-client?
Thanks!
if you press the "generate" button again after the first time it will abort the txt2img. is this what you wanted? progress should be displayed on the webui terminal window if you are using windows, same as running normal webui.
if you press the "generate" button again after the first time it will abort the txt2img. is this what you wanted? progress should be displayed on the webui terminal window if you are using windows, same as running normal webui.
Thanks for the suggestion, but this is not what i am trying to achieve. Also the process currently running is not interrupted if the generate button is pressed again, it stops generating further images. That information on the windows terminal is exaclty what I would like to transfer to aitools-client, that was actually the question if this is passed through the API and available inside aitools-client.
When I originally wrote this there wasn't a way to get the progress status or cancel a render - I believe there is now a a way to do it via api (api/progress endpoint), I'd also like this functionality so maybe I can play with that soon.
PS: About monitoring vram usage via API, not sure, haven't looked into that side of things yet.
