d8b90ac121
big rework of progressbar/preview system to allow multiple users to prompts at the same time and do not get previews of each other
2023-01-15 18:51:04 +03:00
f94a215abe
add an option to choose what you want to see in live preview (Live preview subject) and moves live preview settings to its own tab
2023-01-14 16:29:23 +03:00
42fcc79bd3
add Discard penultimate sigma to infotext
2023-01-05 10:43:21 +03:00
c53852e257
Merge pull request #6044 from hentailord85ez/discard-penultimate-sigma
...
Allow always discarding of penultimate sigma and fix doing 1 less step than specified
2023-01-05 10:33:51 +03:00
bc43293c64
fix incorrect display/calculation for number of steps for hires fix in progress bars
2023-01-04 23:56:43 +03:00
16b9661d27
change karras scheduler sigmas to values recommended by SD from old 0.1 to 10 with an option to revert to old
2023-01-01 09:51:37 +03:00
4df5009acb
Update sd_samplers.py
2022-12-26 20:49:13 +00:00
56e557c6ff
added cheap NN approximation for VAE
2022-12-24 22:39:10 +03:00
a6a54a7529
Merge pull request #5976 from AbstractQbit/fast_preview
...
Add an option for faster low quality previews
2022-12-24 18:38:42 +03:00
0b8acce6a9
separate part of denoiser code into a function to make it easier for extensions to override it
2022-12-24 18:38:16 +03:00
11dd79e346
Add an option for faster low quality previews
2022-12-24 14:00:17 +03:00
399b229783
eliminate duplicated code
...
add an option to samplers for skipping next to last sigma
2022-12-24 09:03:45 +03:00
7ba9bc2fdb
fix dpm2 in img2img as well
2022-12-18 19:16:42 -08:00
180fdf7809
apply to DPM2 (non-ancestral) as well
2022-12-16 08:42:00 -08:00
8b0703b8fc
Add a workaround patch for DPM2 a issue
...
DPM2 a and DPM2 a Karras samplers are both affected by an issue described by https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/3483 and can be resolved by a workaround suggested by the k-diffusion author at https://github.com/crowsonkb/k-diffusion/issues/43#issuecomment-1305195666
2022-12-16 08:18:29 -08:00
0fddb4a1c0
Rework MPS randn fix, add randn_like fix
...
torch.manual_seed() already sets a CPU generator, so there is no reason to create a CPU generator manually. torch.randn_like also needs a MPS fix for k-diffusion, but a torch hijack with randn_like already exists so it can also be used for that.
2022-11-30 10:33:42 -05:00
21effd629d
Add workaround for using MPS with torchsde
2022-11-30 10:33:39 -05:00
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
01f2ed6844
Merge pull request #5065 from JaySmithWpg/vram-leak
...
#3449 - VRAM leak when switching to/from inpainting checkpoint
2022-11-27 13:52:14 +03:00
10923f9b3a
calculate dictionary for sampler names only once
2022-11-27 13:43:10 +03:00
6df4945718
Merge branch 'master' into DPM++SDE
2022-11-27 12:54:45 +03:00
ce6911158b
Add support Stable Diffusion 2.0
2022-11-26 16:10:46 +03:00
c833d5bfaa
fixes #3449 - VRAM leak when switching to/from inpainting model
2022-11-25 20:15:11 -06:00
0a01f50891
Add DPM++ SDE sampler
2022-11-22 14:24:50 +00:00
cdc8020d13
change StableDiffusionProcessing to internally use sampler name instead of sampler index
2022-11-19 12:01:51 +03:00
159475e072
tweak names a bit for new samplers
2022-11-05 18:32:22 +03:00
1b6c2fc749
Reorder samplers
2022-11-04 23:28:13 +00:00
f92dc505a0
Fix name
2022-11-04 23:12:48 +00:00
6008c0773e
Add support for new DPM-Solver++ samplers
2022-11-04 23:03:05 +00:00
eb5e82c7dd
do not unnecessarily run VAE one more time when saving intermediate image with hires fix
2022-11-02 12:45:03 +03:00
9c67408004
Allow saving "before-highres-fix. ( #4150 )
...
* Save image/s before doing highres fix.
2022-11-02 12:18:21 +03:00
10f62546d3
Merge pull request #4021 from AUTOMATIC1111/add-kdiff-cfgdenoiser-callback
...
Add mid-kdiffusion cfgdenoiser script callback - access latents, conditionings and sigmas mid-sampling
2022-11-02 07:29:16 +03:00
5b6bedf6f2
Update class name and assign back to vars
2022-11-02 00:38:17 +00:00
8ae0ea9dea
Add callback to sd_samplers
2022-10-30 23:48:33 +00:00
34c86c12b0
Include PLMS in adjust steps as it also can fail in the same way
2022-10-30 11:04:27 +00:00
de1dc0d279
Add adjust_steps_if_invalid to find next valid step for ddim uniform sampler
2022-10-29 15:23:19 +01:00
b383702752
fix #3145 #3093
2022-10-25 08:32:33 +03:00
d213d6ca6f
removed the option to use 2x more memory when generating previews
...
added an option to always only show one image in previews
removed duplicate code
2022-10-22 20:48:13 +03:00
4fdb53c1e9
Generate grid preview for progress image
2022-10-22 20:36:04 +03:00
d6bd6a425d
Merge branch 'master' into master
2022-10-21 09:49:32 +03:00
d23a46ceaa
Different approach to skip/interrupt with highres fix
2022-10-20 23:49:14 +03:00
92a17a7a4a
Made dummy latents smaller. Minor code cleanups
2022-10-20 09:45:03 -07:00
c418467c03
Don't compute latent mask if were not using it. Also added support for fixed highres_fix generation.
2022-10-19 15:09:43 -07:00
dde9f96072
added support for ddim img2img
2022-10-19 14:14:24 -07:00
8e7097d06a
Added support for RunwayML inpainting model
2022-10-19 13:47:45 -07:00
cbf15edbf9
remove dependence on TQDM for sampler progress/interrupt functionality
2022-10-18 17:23:38 +03:00
eacc03b167
Fix typo in comments
2022-10-11 15:36:29 +03:00
1eae307607
Remove debug code for checking that first sigma value is same after code cleanup
2022-10-11 15:36:29 +03:00
92d7a13885
Handle different parameters for DPM fast & adaptive
2022-10-11 15:36:29 +03:00
7349088d32
--no-half-vae
2022-10-10 16:16:29 +03:00
AUTOMATIC