Commit Graph

81 Commits

Author SHA1 Message Date
Vladimir Mandic 79c682ad4f
fix jpeg 2023-01-04 08:20:42 -05:00
Vladimir Mandic 192ddc04d6
add job info to modules 2023-01-03 10:34:51 -05:00
Vladimir Mandic 1d9dc48efd
init job and add info to model merge 2023-01-03 10:21:51 -05:00
AUTOMATIC 84dd7e8e24 error out with a readable message in chwewckpoint merger for incompatible tensor shapes (ie when trying to merge SD1.5 with SD2.0) 2023-01-02 20:30:02 +03:00
Philpax 6247f21a63 fix(api): don't save extras output to disk 2022-12-24 22:04:53 +11:00
AUTOMATIC1111 ee65237d69
Merge pull request #5747 from yuvalabou/singleton-comparison
Format singleton comparisons
2022-12-24 10:17:38 +03:00
AUTOMATIC1111 55f3ef876b
Merge pull request #5814 from timntorres/5802-save-upscaler-to-filename
Add option to save upscaler to filename suffix in extras tab.
2022-12-24 09:07:00 +03:00
timntorres a7a039d53a Add option to include upscaler name in filename. 2022-12-17 08:50:20 -08:00
timntorres a26fe85056 Add upscaler name as a suffix. 2022-12-17 05:11:06 -08:00
Yuval Aboulafia 957e15c464 Correct singleton comparisons 2022-12-14 20:59:33 +02:00
Mackerel 681c450ecd extras.py: use as little RAM as possible, misc fixes
maximum of 2 models loaded at once. delete unneeded model before next
step. fix 'teritary' -> 'tertiary'. gracefully fail when "add
difference" is selected without a tertiary model
2022-12-04 10:31:06 -05:00
AUTOMATIC 44c46f0ed3 make it possible to merge inpainting model with non-inpainting one 2022-12-04 12:30:44 +03:00
AUTOMATIC 506d529d19 rework #5012 to also work for pictures dragged into the prompt and also add Clip skip + ENSD to parameters 2022-11-27 16:28:32 +03:00
AUTOMATIC1111 488f831d52
Merge pull request #5012 from Nandaka/master
Support NAI style exif in PNG Info for Send... buttons
2022-11-27 15:57:50 +03:00
AUTOMATIC dac9b6f15d add safetensors support for model merging #4869 2022-11-27 15:51:29 +03:00
Nandaka 904121fecc Support NAI exif for PNG Info 2022-11-24 02:39:09 +00:00
Chris OBryan 313e14de04 extras - skip unnecessary second hash of image
There is no need to re-hash the input image each iteration of the loop.
This also reverts PR #4026 as it was determined the cache hits it avoids
were actually valid.
2022-11-02 21:37:43 -05:00
DepFA 29f758afe9
Extend extras image cache with upscale_first arg 2022-10-31 02:39:55 +00:00
AUTOMATIC a1e5e0d766 skip filenames starting with . for img2img and extras batch modes 2022-10-29 08:11:03 +03:00
Chris OBryan d8b3661467 extras: upscaler blending should not be considered in cache key 2022-10-28 16:55:02 -05:00
Chris OBryan 5732c0282d extras-tweaks: autoformat changed lines 2022-10-28 16:36:25 -05:00
Chris OBryan 1f1b327959 extras: Make image cache LRU
This changes the extras image cache into a Least-Recently-Used
cache. This allows more experimentation with different upscalers
without missing the cache.

Max cache size is increased to 5 and is cleared on source image
update.
2022-10-28 16:14:21 -05:00
Chris OBryan bde4731f1d extras: Rework image cache
Bit of a refactor to the image cache to make it easier to extend.
Also takes into account the entire image instead of just a cropped portion.
2022-10-28 14:44:25 -05:00
Chris OBryan 26d0819384 extras: Add option to run upscaling before face fixing
Face restoration can look much better if ran after upscaling, as it
allows the restoration to fix upscaling artifacts. This patch adds
an option to choose which order to run upscaling/face fixing in.
2022-10-28 13:33:49 -05:00
winterspringsummer 9d71eef02e sort file list in alphabetical ordering in extras 2022-10-21 09:00:46 +03:00
winterspringsummer a13c3bed3c Fixed path issue while extras batch processing 2022-10-21 09:00:46 +03:00
winterspringsummer fb5a8cf0d9 Added try except to extras batch from directory 2022-10-21 09:00:45 +03:00
winterspringsummer 60872c5b40 Fixed path issue while extras batch processing 2022-10-21 09:00:45 +03:00
AUTOMATIC ec1924ee57 additional fix for difference model merging 2022-10-18 16:05:52 +03:00
AUTOMATIC e20b7e30fe fix for add difference model merging 2022-10-18 15:33:32 +03:00
DepFA d3338bdef1 extras extend cache key with new upscale to options 2022-10-18 08:29:52 +03:00
AUTOMATIC 5c94aaf290 fix bug for latest model merge RAM improvement 2022-10-17 08:28:18 +03:00
MrCheeze 0fd1307671 improve performance of 3-way merge on machines with not enough ram, by only accessing two of the models at a time 2022-10-17 07:54:36 +03:00
AUTOMATIC 179e3ca752 honor --hide-ui-dir-config option for #2807 2022-10-16 09:51:01 +03:00
winterspringsummer 2ce27728f6 added extras batch work from directory 2022-10-16 09:47:31 +03:00
AUTOMATIC c250cb289c change checkpoint merger to work in a more obvious way
remove sigmoid and inverse sigmoid because they just did the same thing as weighed sum only with changed multiplier
2022-10-14 22:02:32 +03:00
AUTOMATIC 2f0e089c7c should fix the issue with missing layers in chechpoint merger 2022-10-14 21:20:28 +03:00
AUTOMATIC fdecb63685 add an ability to merge three checkpoints 2022-10-14 09:20:24 +03:00
Justin Maier 1d64976dbc Simplify crop logic 2022-10-10 12:04:21 -06:00
Justin Maier 6435691bb1 Add "Scale to" option to Extras 2022-10-09 19:26:52 -06:00
Artem Zagidulin 9ecea0a8d6 fix missing png info when Extras Batch Process 2022-10-09 18:35:25 +03:00
AUTOMATIC bd833409ac additional changes for saving pnginfo for #1803 2022-10-09 13:10:15 +03:00
AUTOMATIC c77c89cc83 make main model loading and model merger use the same code 2022-10-09 10:23:31 +03:00
Jairo Correa ad1fbbae93 Merge branch 'master' into fix-vram 2022-09-30 18:58:51 -03:00
AUTOMATIC 3f64e23f81 fix model checkpoint merger tab failing to save model 2022-09-30 22:57:25 +03:00
AUTOMATIC d1f098540a remove unwanted formatting/functionality from the PR 2022-09-30 11:42:40 +03:00
d8ahazard d73741794d Merge remote-tracking branch 'upstream/master' into ModelLoader 2022-09-29 19:59:36 -05:00
d8ahazard 0dce0df1ee Holy $hit.
Yep.

Fix gfpgan_model_arch requirement(s).
Add Upscaler base class, move from images.
Add a lot of methods to Upscaler.
Re-work all the child upscalers to be proper classes.
Add BSRGAN scaler.
Add ldsr_model_arch class, removing the dependency for another repo that just uses regular latent-diffusion stuff.
Add one universal method that will always find and load new upscaler models without having to add new "setup_model" calls. Still need to add command line params, but that could probably be automated.
Add a "self.scale" property to all Upscalers so the scalers themselves can do "things" in response to the requested upscaling size.
Ensure LDSR doesn't get stuck in a longer loop of "upscale/downscale/upscale" as we try to reach the target upscale size.
Add typehints for IDE sanity.
PEP-8 improvements.
Moar.
2022-09-29 17:46:23 -05:00
safentisAuth 66fed8ffb8 Add custom name and try-except 2022-09-29 12:41:47 +03:00
Jairo Correa c938679de7 Fix memory leak and reduce memory usage 2022-09-28 22:14:13 -03:00