Twanzito issues

Repositories
Issues
Comments

Results 2 issues of


                                            Twanzito

Loading lora weights in pipeline.py has no effect on the models.

Also, is there a way to use a non coreML text-encoder or to make this compatible with Compel?

build / option to switch torch device from cuda to mps for M2/M1 Apple Silicon (./deps/ebsynth/build-macos-mps_only.sh) ?

Has this been tested? https://lightning.ai/docs/pytorch/stable/accelerators/mps_basic.html About to regex out the to('cuda') and device='cuda', and turn off fps16 and give it a try