Bryce
f97f6a3b4b
feature: use refiners library for generation
...
BREAKING CHANGE
- stable diffusion 1.5 + inpainting working
- self-attention guidance working. improves image generation quality
- tile-mode working
- inpainting self-attention guidance working
disable/broken features:
- sd 1.4, 2.0, 2.1
- most of the samplers
- pix2pix edit
- most of the controlnets
- memory management
- python 3.8 support
wip
9 months ago
Bryce
558d3388e5
style: speed up linting and autoformatting. fix lints
11 months ago
Bryce
d32e1060cd
feature: multi-controlnet support at the command line
...
add controlnet option for edit demo
1 year ago
Bryce
bcaa000d35
fix: model logging
1 year ago
Bryce
926692ad03
tests: "prime" the controlnets
...
Trying to get things working on m1. doesn't fix everything
1 year ago
Bryce
eca97a25a0
tests: adjust tests to pass
1 year ago
Bryce
6db296aa37
tests: faster tests
1 year ago
Bryce
3b777b98d8
fix: improved large images using composition
2 years ago
Bryce
70c58467c0
fix: img2img was broken for all samplers except ddim,plms
...
img2img was broken for all samplers except plms and ddim when init image strength was >~0.25. Been this way for a while. whoops
2 years ago
Bryce
54c3ad51d6
feature: controlnet
2 years ago
Bryce
003a512dc8
perf: mps perf improvement
...
attention slicing wasn't working
2 years ago
Bryce
ea1d4baafe
refactor: cleanup image generation code
2 years ago
Bryce
8a97213622
feature: sliced latent decoding
...
allows generation of bigger images. tile seams can be noticeable occasionally despite the feathering
2 years ago
Bryce Drennan
542e4fbd55
fix: tile mode was broken since latest perf improvements ( #220 )
2 years ago
Bryce
9e0a9e2c29
tests: update
2 years ago
Bryce
1381c7fed4
lint: new ruff linter
2 years ago
Bryce
da0f1e1ee6
fix: make sure randomness is generated on cpu for consistency
2 years ago
Bryce
1f7403155e
fix: more thorough cleaning of memory when switching models
...
also cleanup up some test failures
2 years ago
Bryce
f30823e0b5
feature: Stable Diffusion 2.1
2 years ago
Bryce
9c153a8d2d
feature: support k-sampler methods for SD-2.0-v (768)
2 years ago
Bryce
015088507f
feature/refactor/fix: better defaults. correct version in metadata
...
- feature: use different default steps and image sizes depending on sampler and model selceted
- fix : #110 use proper version in image metadata
- refactor: samplers all have their own class that inherits from ImageSampler
2 years ago
Bryce
893b041a8f
feature: progress image callback
2 years ago
Bryce
94c0aeab12
feature: switch between models without restarting python instance
2 years ago
Bryce
8332593fed
feature: dilation and erosion of masks
...
Previously the `+` and `-` characters in a mask (example: `face{+0.1}`) added to the grayscale value of any masked areas. This wasn't very useful. The new behavior is that the mask will expand or contract by the number of pixel specified. The technical terms for this are dilation and erosion. This allows much greater control over the masked area.
2 years ago
Bryce
741a433c56
feature: update k_diffusion. add dpm_fast and dpm_addaptive samplers
2 years ago
Bryce
dcf953383e
tests: support distributed test runs
...
- switch to expected images instead of hashes. allow fuzzy matches
feature: more consistent seeds
2 years ago
Bryce
4ba1965db8
feature: k-diff sampler img2img and masking
2 years ago
jsdman1313@gmail.com
740870ad8e
fix: updates cats to dogs test to check between two different hashes
2 years ago
Bryce
281d23adf2
refactor: prep for supporting img2img in k-samplers
2 years ago
Bryce
db5ebe9095
tests: tests can run without any network calls
2 years ago
Bryce
f21c979f08
feature: better masking segmentation from clipseg
2 years ago
Bryce
bc135724a3
fix: fix memory leak in face enhancer
...
thanks to @h4rk8s for discovering and finding a remedy
root cause was a model being instantiated inside
FaceRestoreHelper
2 years ago
Bryce
95a8fa31a9
fix: inpainting producing blurry images
...
while the previous version did produce much better blending it also makes images that lack detail for some reason.
tests: Added more tests to help catch this sort of thing earlies
fix: found that median blur is really slow, so I made sure we only do it on downsampled masks. Was taking like 3 minutes to run on the large pearl girl picture on M1
- docs: update examples
2 years ago