Environment
I want to just show demo with using WebUI. So I need to just WebUI working.
COMMANDLINE_ARGS="--skip-torch-cuda-test --upcast-sampling --no-half-vae --use-cpu interrogate --precision full --no-half --opt-split-attention-v1"
I'm using arguments above. And It seems like working well at moment, but the result is very bad. The example here.
When I removed the argument, "--opt-split-attention-v1", another error was made.
RuntimeError: MPS backend out of memory (MPS allocated: 5.05 GB, other allocations: 2.24 GB, max allowed: 6.77 GB). Tried to allocate 1024.00 MB on private pool. Use PYTORCH_MPS_HIGH_WATERMARK_RATIO=0.0 to disable upper limit for memory allocations (may cause system failure).
I guess that the graphic card memory is too lower to run WebUI by my mac. But using DiffusionBee, I can make new image by same mechanism.
Why I can't make image by WebUI? What is difference between WebUI and DiffussionBee?
Ok, I found the answer. I just add argument --use-cpu all with --no-half.
And I remove argument --opt-split-attention-v1.
It takes longer time to generate new image, but WebUI is working.
