You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
ggml_extend.hpp:1000 - vae compute buffer size: 2496.00 MB(VRAM)
Native API failed. Native API returns: 40 (UR_RESULT_ERROR_OUT_OF_RESOURCES)
This happens on the second run of stable-diffusion-cpp-python.
First "run" with as many generations as needed completes succesfully.
Not sure if it's my fault that i should destroy the pipeline somehow and not just let the script end.
Alternately probably a bug of intel drivers (i'm on 22.04, i could try to upgrade to 24.04 and see if that makes any diff)
ggml_extend.hpp:1000 - vae compute buffer size: 2496.00 MB(VRAM)
Native API failed. Native API returns: 40 (UR_RESULT_ERROR_OUT_OF_RESOURCES)
This happens on the second run of stable-diffusion-cpp-python.
First "run" with as many generations as needed completes succesfully.
Not sure if it's my fault that i should destroy the pipeline somehow and not just let the script end.
Alternately probably a bug of intel drivers (i'm on 22.04, i could try to upgrade to 24.04 and see if that makes any diff)
using 0.2.1 version of the bindings. 0.2.2 do not work due to a different bug (leejet/stable-diffusion.cpp#542)
The text was updated successfully, but these errors were encountered: