Commit Graph

56 Commits

Author SHA1 Message Date
AUTOMATIC1111 9553a7e071
Merge pull request #3818 from jwatzman/master
Reduce peak memory usage when changing models
2022-10-29 09:16:00 +03:00
Antonio 5d5dc64064
Natural sorting for dropdown checkpoint list
Example:

Before					After

11.ckpt					11.ckpt
ab.ckpt					ab.ckpt
ade_pablo_step_1000.ckpt	ade_pablo_step_500.ckpt			
ade_pablo_step_500.ckpt	ade_pablo_step_1000.ckpt	
ade_step_1000.ckpt		ade_step_500.ckpt
ade_step_1500.ckpt		ade_step_1000.ckpt
ade_step_2000.ckpt		ade_step_1500.ckpt
ade_step_2500.ckpt		ade_step_2000.ckpt
ade_step_3000.ckpt		ade_step_2500.ckpt
ade_step_500.ckpt			ade_step_3000.ckpt
atp_step_5500.ckpt			atp_step_5500.ckpt
model1.ckpt				model1.ckpt
model10.ckpt				model10.ckpt
model1000.ckpt			model33.ckpt
model33.ckpt				model50.ckpt
model400.ckpt			model400.ckpt
model50.ckpt				model1000.ckpt
moo44.ckpt				moo44.ckpt
v1-4-pruned-emaonly.ckpt	v1-4-pruned-emaonly.ckpt
v1-5-pruned-emaonly.ckpt	v1-5-pruned-emaonly.ckpt
v1-5-pruned.ckpt			v1-5-pruned.ckpt
v1-5-vae.ckpt				v1-5-vae.ckpt
2022-10-28 05:49:39 +02:00
Josh Watzman b50ff4f4e4 Reduce peak memory usage when changing models
A few tweaks to reduce peak memory usage, the biggest being that if we
aren't using the checkpoint cache, we shouldn't duplicate the model
state dict just to immediately throw it away.

On my machine with 16GB of RAM, this change means I can typically change
models, whereas before it would typically OOM.
2022-10-27 22:01:06 +01:00
AUTOMATIC 321bacc6a9 call model_loaded_callback after setting shared.sd_model in case scripts refer to it using that 2022-10-22 20:15:12 +03:00
MrCheeze 0df94d3fcf fix aesthetic gradients doing nothing after loading a different model 2022-10-22 20:14:18 +03:00
AUTOMATIC 2b91251637 removed aesthetic gradients as built-in
added support for extensions
2022-10-22 12:23:58 +03:00
AUTOMATIC ac0aa2b18e loading SD VAE, see PR #3303 2022-10-21 17:35:51 +03:00
AUTOMATIC df57064093 do not load aesthetic clip model until it's needed
add refresh button for aesthetic embeddings
add aesthetic params to images' infotext
2022-10-21 16:10:51 +03:00
AUTOMATIC 7d6b388d71 Merge branch 'ae' 2022-10-21 13:35:01 +03:00
random_thoughtss 49533eed9e XY grid correctly re-assignes model when config changes 2022-10-20 16:01:27 -07:00
random_thoughtss 708c3a7bd8 Added PLMS hijack and made sure to always replace methods 2022-10-20 13:28:43 -07:00
random_thoughtss 8e7097d06a Added support for RunwayML inpainting model 2022-10-19 13:47:45 -07:00
AUTOMATIC f894dd552f fix for broken checkpoint merger 2022-10-19 12:45:42 +03:00
MalumaDev 2362d5f00e
Merge branch 'master' into test_resolve_conflicts 2022-10-19 10:22:39 +02:00
AUTOMATIC 10aca1ca3e more careful loading of model weights (eliminates some issues with checkpoints that have weird cond_stage_model layer names) 2022-10-19 08:42:22 +03:00
MalumaDev 9324cdaa31 ui fix, re organization of the code 2022-10-16 17:53:56 +02:00
AUTOMATIC1111 af144ebdc7
Merge branch 'master' into ckpt-cache 2022-10-15 10:35:18 +03:00
Rae Fu e21f01f645 add checkpoint cache option to UI for faster model switching
switching time reduced from ~1500ms to ~280ms
2022-10-14 14:09:23 -06:00
AUTOMATIC bb295f5478 rework the code for lowram a bit 2022-10-14 20:03:41 +03:00
Ljzd-PRO 4a216ded43 load models to VRAM when using `--lowram` param
load models to VRM instead of RAM (for machines which have bigger VRM than RAM such as free Google Colab server)
2022-10-14 19:57:23 +03:00
AUTOMATIC 727e4d1086 no to different messages plus fix using != to compare to None 2022-10-10 20:46:55 +03:00
AUTOMATIC1111 b3d3b335cf
Merge pull request #2131 from ssysm/upstream-master
Add VAE Path Arguments
2022-10-10 20:45:14 +03:00
ssysm af62ad4d25 change vae loading method 2022-10-10 13:25:28 -04:00
AUTOMATIC 7349088d32 --no-half-vae 2022-10-10 16:16:29 +03:00
ssysm 6fdad291bd Merge branch 'master' of https://github.com/AUTOMATIC1111/stable-diffusion-webui into upstream-master 2022-10-09 23:20:39 -04:00
ssysm cc92dc1f8d add vae path args 2022-10-09 23:17:29 -04:00
AUTOMATIC e6e8cabe0c change up #2056 to make it work how i want it to plus make xy plot write correct values to images 2022-10-09 14:57:48 +03:00
William Moorehouse d6d10a37bf Added extended model details to infotext 2022-10-09 14:49:15 +03:00
AUTOMATIC f4578b343d fix model switching not working properly if there is a different yaml config 2022-10-09 13:23:30 +03:00
AUTOMATIC 4e569fd888 fixed incorrect message about loading config; thanks anon! 2022-10-09 10:31:47 +03:00
AUTOMATIC c77c89cc83 make main model loading and model merger use the same code 2022-10-09 10:23:31 +03:00
AUTOMATIC 050a6a798c support loading .yaml config with same name as model
support EMA weights in processing (????)
2022-10-08 23:26:48 +03:00
Aidan Holland 432782163a chore: Fix typos 2022-10-08 22:42:30 +03:00
leko 616b7218f7 fix: handles when state_dict does not exist 2022-10-08 12:38:50 +03:00
AUTOMATIC d15b3ec001 support loading VAE 2022-10-07 10:40:22 +03:00
AUTOMATIC 852fd90c0d emergency fix for disabling SD model download after multiple complaints 2022-10-02 21:22:20 +03:00
AUTOMATIC a1cde7e646 disabled SD model download after multiple complaints 2022-10-02 21:09:10 +03:00
AUTOMATIC 0758f6e641 fix --ckpt option breaking model selection 2022-10-02 17:24:50 +03:00
AUTOMATIC 820f1dc96b initial support for training textual inversion 2022-10-02 15:03:39 +03:00
AUTOMATIC 2b03f0bbda if --ckpt option is specified, load that model 2022-09-30 22:16:03 +03:00
AUTOMATIC cef838a6ab revert the annotation not supported by old pythons 2022-09-30 12:15:29 +03:00
AUTOMATIC d1f098540a remove unwanted formatting/functionality from the PR 2022-09-30 11:42:40 +03:00
AUTOMATIC 8f1b315318 fix bugs in the PR 2022-09-30 09:46:52 +03:00
AUTOMATIC1111 25414bcd05
Merge pull request #1109 from d8ahazard/ModelLoader
Model Loader, Fixes
2022-09-30 09:35:58 +03:00
DepFA ebd2c48115 return shortest checkpoint title match 2022-09-30 07:37:05 +03:00
DepFA 642b7e333e add get_closet_checkpoint_match 2022-09-30 07:37:05 +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
AUTOMATIC c715ef04d1 fix for incorrect model weight loading for #814 2022-09-29 15:40:28 +03:00
AUTOMATIC 29ce8a687d remove unneded debug print 2022-09-29 08:03:23 +03:00