Damian at mba
d572af2acf
fix cross-attention on k* samplers
2022-10-19 21:08:03 +02:00
Damian at mba
54e6a68acb
wip bringing cross-attention to PLMS and DDIM
2022-10-19 21:08:03 +02:00
Damian at mba
09f62032ec
cleanup and clarify comments
2022-10-19 21:08:03 +02:00
Damian at mba
711ffd238f
cleanup
2022-10-19 21:08:03 +02:00
Damian at mba
056cb0d8a8
sliced cross-attention wrangler works
2022-10-19 21:08:03 +02:00
Damian at mba
37a204324b
go back to using InvokeAI attention
2022-10-19 21:08:03 +02:00
Damian at mba
1fc1f8bf05
cross-attention working with placeholder {} syntax
2022-10-19 21:06:42 +02:00
Damian at mba
8ff507b03b
runs but doesn't work properly - see below for test prompt
...
test prompt:
"a cat sitting on a car {a dog sitting on a car}" -W 384 -H 256 -s 10 -S 12346 -A k_euler
note that substition of dog for cat is currently hard-coded (ksampler.py
line 43-44)
2022-10-19 21:06:42 +02:00
Damian at mba
33d6603fef
cleanup initial experiments
2022-10-19 21:06:42 +02:00
Damian at mba
b0b1993918
initial experiments
2022-10-19 21:06:42 +02:00
Lincoln Stein
e98fe9c22d
fix noisy images at high step counts
...
At step counts greater than ~75, the ksamplers start producing noisy
images when using the Karras noise schedule. This PR reverts to using
the model's own noise schedule, which eliminates the problem at the
cost of slowing convergence at lower step counts.
This PR also introduces a new CLI `--save_intermediates <n>' argument,
which will save every nth intermediate image into a subdirectory
named `intermediates/<image_prefix>'.
Addresses issue #1083 .
2022-10-14 00:01:59 -04:00
Lincoln Stein
c5344acb25
run make_schedule() if it hasn't already been called
...
- fixes #984
2022-10-09 15:30:23 -04:00
Lincoln Stein
8faa06fb15
Merge branch 'main' into development
...
- this syncs documentation and code
2022-10-09 14:47:27 -04:00
Lincoln Stein
2b1aaf4ee7
rename all modules from ldm.dream to ldm.invoke
...
- scripts and documentation updated to match
- ran preflight checks on both web and CLI and seems to be working
2022-10-08 11:37:23 -04:00
Lincoln Stein
7a701506a4
restore ability of ksamplers to process -v variation options
...
- supersedes PR #977
- works with both img2img and txt2img
2022-10-07 16:25:58 -04:00
Lincoln Stein
5157cbeda1
restore ability of ksamplers to process -v variation options
...
- supersedes #977
2022-10-07 16:21:16 -04:00
Lincoln Stein
9f34ddfcea
fix crash on len(Nonetype) in k_sampler
2022-10-07 08:05:13 -04:00
Lincoln Stein
d60df54f69
fix k_samplers in img2img - probably correct now
2022-10-06 18:53:54 -04:00
Lincoln Stein
7541c7cf5d
fix k_samplers in img2img - probably correct now
2022-10-06 18:31:04 -04:00
Lincoln Stein
440065f7f8
revert previous change
2022-10-06 14:57:06 -04:00
Lincoln Stein
2c27e759cd
fix #889 - fuzzy k* img2img at low strength
2022-10-06 14:16:56 -04:00
Lincoln Stein
f3050fefce
bug and warning message fixes
...
- txt2img2img back to using DDIM as img2img sampler; results produced
by some k* samplers are just not reliable enough for good user
experience
- img2img progress message clarifies why img2img steps taken != steps requested
- warn of potential problems when user tries to run img2img on a small init image
2022-10-06 10:39:08 -04:00
Peter Baylies
5d911b43c0
Merge branch 'development' into development
2022-10-04 18:40:42 -04:00
Lincoln Stein
483097f31c
fix off-by-one error
2022-10-05 00:15:47 +02:00
Lincoln Stein
7a3eae4572
revert to original k* noise schedule
2022-10-05 00:15:47 +02:00
Peter Baylies
e79069a957
Merge branch 'development' into development
2022-10-03 08:33:44 -04:00
Lincoln Stein
8e97bc24a4
restore step argument to step_callback
2022-10-03 05:38:43 -04:00
Lincoln Stein
a0f4af087c
restore use of sampler.decode() in img2img
2022-10-02 14:58:21 -04:00
Lincoln Stein
958d7650dd
img2img works with all samplers, inpainting working with ddim & plms
...
- img2img confirmed working with all samplers
- inpainting working on ddim & plms. Changes to k-diffusion
module seem to be needed for inpainting support.
- switched k-diffuser noise schedule to original karras schedule,
which reduces the step number needed for good results
2022-10-02 14:58:21 -04:00
Peter Baylies
dc500946ad
* Make warmup feature less aggressive to support higher thresholding values properly.
2022-09-07 04:50:53 -04:00
Peter Baylies
0891910cac
* Updates for thresholding and perlin noise options, added warmup for thresholding.
2022-09-05 21:40:05 -04:00
Peter Baylies
1a4bed2e55
Merge branch 'development' of https://github.com/lstein/stable-diffusion into development
2022-09-03 05:02:25 -04:00
Peter Baylies
b6cf8b9052
* Add threshold and perlin noise options for Karras samplers.
2022-09-02 13:39:26 -04:00
Kevin Gibbons
4fe265735a
support generating variations
...
Co-authored-by: xra <mail@xra.dev>
2022-09-01 23:48:53 -07:00
Lincoln Stein
bdb0651eb2
add support for Apple hardware using MPS acceleration
2022-08-31 00:33:23 -04:00
tesseractcat
070795a3b4
webui: stream progress events to page
2022-08-28 21:54:10 -07:00
Lincoln Stein
4f02b72c9c
prettified all the code using "blue" at the urging of @tildebyte
2022-08-26 03:15:42 -04:00
BlueAmulet
b7735b3788
Fix attribution
2022-08-25 11:13:12 -06:00
BlueAmulet
39b55ae016
Remove accelerate library
...
This library is not required to use k-diffusion
Make k-diffusion wrapper closer to the other samplers
2022-08-25 11:04:57 -06:00
Ben Alkov
050dffd269
feat(samplers): add ability use all k_* samplers
...
Signed-off-by: Ben Alkov <ben.alkov@gmail.com>
2022-08-23 17:26:22 -04:00
Lincoln Stein
78aba5b770
preparing for merge into main
2022-08-21 19:57:48 -04:00
Lincoln Stein
bb91ca0462
first attempt to fold k_lms changes proposed by hwharrison and bmaltais
2022-08-21 17:09:00 -04:00
