- 13 Aug, 2023 37 commits
-
-
brkirch authored
-
brkirch authored
-
brkirch authored
-
brkirch authored
For MPS, using a tensor created with `torch.empty()` can cause `torch.baddbmm()` to include NaNs in the tensor it returns, even though `beta=0`. However, with a tensor of shape [1,1,1], there should be a negligible performance difference between `torch.empty()` and `torch.zeros()` anyway, so it's better to just use `torch.zeros()` for this and avoid unnecessarily creating issues.
-
brkirch authored
-
brkirch authored
Even if this causes chunks to be much smaller, performance isn't significantly impacted. This will usually reduce memory usage but should also help with poor performance when free memory is low.
-
AUTOMATIC1111 authored
-
AUTOMATIC1111 authored
Support search and display of hashes for all extra network items
-
AUTOMATIC1111 authored
Clear sampler and garbage collect before decoding images to reduce VRAM
-
AUTOMATIC1111 authored
Encode batch items individually to significantly reduce VRAM
-
AUTOMATIC1111 authored
Update description of eta setting
-
AUTOMATIC1111 authored
Add `s_noise` param to more samplers
-
AUTOMATIC1111 authored
Restore `extra_params` that was lost in merge
-
AUTOMATIC1111 authored
-
catboxanon authored
-
catboxanon authored
-
catboxanon authored
-
catboxanon authored
-
catboxanon authored
-
AUTOMATIC1111 authored
Lora: output warnings in UI rather than fail for unfitting loras; switch to logging for error output in console
-
catboxanon authored
-
catboxanon authored
-
catboxanon authored
More significant VRAM reduction.
-
catboxanon authored
Significantly reduces VRAM. This makes encoding more inline with how decoding currently functions.
-
catboxanon authored
-
AUTOMATIC1111 authored
Add Norm Module to lora ext and add "bias" support
-
AUTOMATIC1111 authored
Bring back CSV mode for XYZ grid
-
AUTOMATIC1111 authored
-
w-e-w authored
-
AUTOMATIC1111 authored
write out correct model name in infotext, rather than the refiner model
-
Kohaku-Blueleaf authored
-
Kohaku-Blueleaf authored
-
Kohaku-Blueleaf authored
refer to LyCORIS 1.9.0.dev6 add new option and module for training norm layer (Which is reported to be good for style)
-
w-e-w authored
-
w-e-w authored
-
w-e-w authored
-
AUTOMATIC1111 authored
-
- 12 Aug, 2023 3 commits
-
-
AUTOMATIC1111 authored
-
AUTOMATIC1111 authored
-
AUTOMATIC1111 authored
-