Commit Graph

519 Commits (fbb16f6c62a91737b32f685b23d2a7b647a3ca93)
 

Author SHA1 Message Date
Bryce fbb16f6c62 version: 14.1.1 4 months ago
Bryce 279dc28b1d build: triton only has wheels for linux 4 months ago
Bryce 3906072191 ci: test installation on windows, mac, and conda 4 months ago
Bryce 60baed839c build: remove some dependency version limits 4 months ago
Bryce 4639dfb041 version: 4.1.0 4 months ago
Bryce Drennan 1de5d28554
test: run some tests on python 3.11 as well (#454)
torch doesn't support 3.12 yet
5 months ago
Bryce Drennan 601a112dc3
refactor: move download related functions to separate module (#453)
+ renames and typehints
5 months ago
Bryce Drennan 502ffbdc63
feature: sdxl inpaint support (#450) 5 months ago
Bryce Drennan 700cb457b9
feature: support loading sdxl compvis weights (#449) 5 months ago
Bryce Drennan 907e80d1f2
feature: video interpolation (#448)
- uses rife algorithm to interpolate frames
5 months ago
Bryce Drennan bb2dd45cf2
feature: videogen improvements (#447)
- allow generation at any size
- output "bounce" animations
- choose output format: mp4, webp, or gif
- fix random seed handling
- organize some code better
5 months ago
Bryce Drennan c5199ed7cc
docs: minor fixes (#446) 5 months ago
Bryce Drennan 66b17cc315
docs: add github action to push docs (#444) 5 months ago
Bryce 807c976da3 build: remove imageio dependency 5 months ago
Bryce d2609cb5cd fix: use smaller composition size 5 months ago
Bryce 5bbb09f69e build: vendorize facexlib
had too many unused sub-dependencies

also monkeypatch the download mechanism to use our standard download function
5 months ago
Bryce 4521d518ac version: 14.0.4 5 months ago
Bryce Drennan d3106fc9e3
fix: videogen bug (#443) 5 months ago
jaydrennan 89bc1a9f1c
docs: adds docs tool, material for mkdocs, along with more fleshed ou… (#428)
* docs: adds docs tool, material for mkdocs, along with more fleshed out docstrings.

this includes ability to serve up a local docs website.


---------

Co-authored-by: Bryce <github20210803@accounts.brycedrennan.com>
5 months ago
Bryce Drennan 0271bffa38
build: remove fairscale dependency (#441) 5 months ago
Bryce 12e4855792 version: 14.0.3 5 months ago
Bryce 7ea4cba5d3 fix: add missing dependency. add package smoketest 5 months ago
Bryce f866aaa44d version: 14.0.2 5 months ago
Bryce e00c7b9eb7 fix: add back missing init file 5 months ago
Bryce 1598469179 version: 14.0.1 5 months ago
Bryce d148bc1537 fix: progress latent collection bug 5 months ago
Bryce fdb48399af docs: separate changelog. other small cleanup 5 months ago
Bryce ed40a12c01 version: 14.0.0 5 months ago
Bryce 57dc27df8c build: tag imaginairy as typed 5 months ago
Bryce 55e27160f5 build: vendorize refiners
so we can still work in conda envs
5 months ago
Bryce f84406f12c fix: handle unexpected keys in weights better 5 months ago
Bryce 5b3b04b877 build: remove pytorch lightning dependency 5 months ago
Bryce 9f33fa0664 version: 14.0.0b9 5 months ago
Bryce 35a5ccbcba fix: include weightmaps 5 months ago
Bryce e90bf1b47d version: 14.0.0b8 5 months ago
Bryce 7100d3f9ea perf: make upscaler use fp16 for better efficiency 5 months ago
Bryce 4fcfc363af fix: always show total last 5 months ago
Bryce f50a1f5b0c fix: interrupted generations don't prevent more generations
fixes #424

- pref: improve memory usage when loading SD15.
- feature: clean up CLI output more
- feature: cuda memory tracking context manager
- feature: use safetensors fp16 for sd15
5 months ago
Bryce 9e3403df89 feature: clean up terminal output
- recording timing and memory usage of various steps
- re-use logging context for composition images
- load sdxl weights in a more VRAM efficient way
- switch to diffusers weights for default weights for sd15
5 months ago
Bryce 0d78b8271f version: 14.0.0b7 5 months ago
Bryce Drennan 77c4b85037
perf: improve memory usage (#433)
add warning for corrupt weights files
5 months ago
Bryce 26d1ff9bc4 version: 14.0.0b6 5 months ago
Bryce Drennan 42a045e8e6
feature: support sdxl (#431)
- adds support for (SDXL)[https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0]
  - adds sliced encoding/decoding to refiners sdxl pipeline
  - doesn't support inpainting, controlnets
- monkeypatches self_attention_guidance to use sliced attention
- adds a bunch of model weight translation utilities and weightmaps
- add [opendalle 1.1](https://huggingface.co/dataautogpt3/OpenDalleV1.1)
- change default model to opendalle
- fix: better handle special characters in path inputs on command line
**todo**
- add tests
5 months ago
jaydrennan fb7cff3684
Merge pull request #430 from brycedrennan/ruff_formatter
Ruff formatter
5 months ago
jaydrennan 3322777f5e refactor: formatting changes ruff formatter
ruff formatter has slight differences in formatting compared to black
5 months ago
jaydrennan 7eef3bf628 feature: replaces black formatter with ruff formatter 5 months ago
Bryce Drennan a2c38b3ec0
feature: support loading diffusers folder/based models from huggingface (#427) 5 months ago
Bryce 50e796a3b7 refactor: move code around 5 months ago
Bryce 32b5175e0e feature: better upscaling
- use face enhancement in a smarter way that doesn't blur high-res images
- use a different upscale model for composition images

**Upscaling**
RealESRGAN is great but it blurs parts of images it doesn't understand

4xUltrasharp is a finetune of RealESRGan that isn't as good but doesn't have this blurry patch problem.  This makes it more suitable to use as part of the composition/upscale process.  We still use realesrgan for any last-step upscales since it does look better.

had to write a state dict translator to use the ultrasharp model

**Face Enhancement**

We no longer enhance faces that are larger than 512 pixels. They should already have enough details and the face enhancer doesn't produce faces at high enough resolution to look good at that size.
5 months ago
Bryce 6ebd12abb1 refactor: move code to more intuitive places 5 months ago