333e2bb30b
fix for qwen image crashing due to ref images being too big, trial and error shows it happens after 512x512
2025-11-02 01:31:01 +08:00
7179e49aef
fix from https://github.com/leejet/stable-diffusion.cpp/pull/926
2025-11-01 23:38:37 +08:00
0891b0752d
qwen3vl fixed (+2 squashed commit)
...
Squashed commit:
[89f65ed0c] wip fixing q3vl
[6fa34cff2] wip fixing q3vl
2025-10-31 17:52:33 +08:00
c2316353a1
allow usage of flux without some components
2025-10-30 22:32:20 +08:00
96a70033ba
sd: sync to master-343-dd75fc0 ( #1818 )
2025-10-30 13:44:59 +08:00
c652d08f02
sd: sync to master-340-9e28be6 ( #1816 )
2025-10-27 21:47:48 +08:00
2cab657c60
sd: sync to master-336-917f7bf ( #1810 )
2025-10-25 21:19:35 +08:00
97867f1990
add alt umt5xxl tensor name ( #1813 )
2025-10-25 12:48:08 +08:00
fef73919ea
sd: clean up changes against stable-diffusion.cpp 90ef5f8 ( #1804 )
...
* sd: clean up changes against stable-diffusion.cpp 90ef5f8
Clean up the diff, and include a few missing changes, mainly from
the upscaler and model weight type statistics.
* added line clear again
* remove excess spaces
---------
Co-authored-by: LostRuins Concedo <39025047+LostRuins@users.noreply.github.com >
2025-10-23 22:00:33 +08:00
86b94456de
Sync sd.cpp to 90ef5f8
2025-10-20 11:07:42 +08:00
fab2ff0687
sync sd.cpp to e370258
2025-10-20 10:45:34 +08:00
49535fdfed
rename define
2025-10-20 10:08:39 +08:00
d7da1eb35c
invert KCPP_BAKE_SD_VOCAB logic, move define to sdtype_adapter.cpp ( #1803 )
...
Using KCPP_BAKE_SD_VOCAB to turn off the change to not embed the
vocabulary files makes testing new upstream merges harder, because
we then need to set that macro on the sd.cpp original build.
So, revert the tests, making the define turn the change on. Also,
since model.cpp is always built by Koboldcpp as part of the
sdtype_adapter.cpp, it's enough to set the macro on that file.
2025-10-20 10:07:37 +08:00
d9e898afe0
reset scheduler if default otherwise it will persist the old one
2025-10-19 19:53:40 +08:00
f6916ba864
updated sdui
2025-10-17 13:56:45 +08:00
c18d7991c8
rename tts.cpp ggml_round to ggml_ttsround to avoid conflict
2025-10-16 14:24:44 +08:00
5207b8d4be
more sd path fallbacks
2025-10-15 15:22:06 +08:00
610ba18971
sdcpp precision fix
2025-10-15 11:08:35 +08:00
54db35cd7a
fix t5 scale as well
2025-10-12 20:35:46 +08:00
1a360b8458
sdcpp: optimize the handling of the FeedForward precision fix (+1 squashed commits)
...
Squashed commits:
[621ff6392] sdcpp: optimize the handling of the FeedForward precision fix (+1 squashed commits)
Squashed commits:
[05b16906c] sdcpp: optimize the handling of the FeedForward precision fix
2025-10-12 17:49:38 +08:00
76d5fcbe49
fix the issue that occurs when using CUDA with k-quants weights
2025-10-12 16:18:03 +08:00
65c2129f65
https://github.com/leejet/stable-diffusion.cpp/pull/877/commits/47c0f8e4bd6916442d04b0a4412554cf3a043e8d
2025-10-12 10:01:29 +08:00
0cc0ea4cf9
reset prompt template idx
2025-10-11 12:30:07 +08:00
5cea2fe944
don't enforce dims
2025-10-11 11:34:47 +08:00
80f88eb703
wip qwen image edit. not working yet
2025-10-11 11:24:17 +08:00
bc09f34f66
only accept qwen image pruned models matching 40 or 41 layers
2025-10-10 16:26:32 +08:00
bc762fe9b4
add support for Qwen Image Pruning ( #1779 )
...
From leejet/stable-diffusion.cpp#874 .
2025-10-10 16:22:47 +08:00
bece22f996
fix encoding VAE tiling for Qwen Image ( #1785 )
2025-10-10 10:07:50 +08:00
c48999f7c0
additional options for image generation ( #1765 )
...
* sd: add backend support for choosing the default sampler
* use the default sampler on the API
* sd: add backend support for the scheduler
* sd: add backend support for distilled guidance
* sd: add backend support for timestep-shift
* sd: add a config field to set default image gen options
2025-10-05 23:36:20 +08:00
75272f62af
remove gif-h
2025-10-05 17:49:29 +08:00
a98b63013e
allow tiling on qwen image
2025-10-04 20:43:36 +08:00
bb06956b2d
allow wan to use img2img via init image
2025-10-04 11:25:46 +08:00
db37688b47
qwen image disable VAE tiling as it's broken
2025-10-04 11:19:19 +08:00
7447a362d1
hide a debug print
2025-10-03 20:56:57 +08:00
15249baea1
apply jeffbolz f32 patch https://github.com/leejet/stable-diffusion.cpp/pull/851#issuecomment-3335515302
2025-10-03 19:18:46 +08:00
f282362414
added qwen image support (+1 squashed commits)
...
Squashed commits:
[92df28061] added qwen image support (+1 squashed commits)
Squashed commits:
[1485c71ed] wip adding qwen image
2025-10-03 18:58:48 +08:00
4f8f0e5949
move embeds into their own dir, detach sd vocab into separate files
2025-10-03 14:21:09 +08:00
f2ad0b78d4
refactor of added images done
2025-10-02 21:22:35 +08:00
df87da4694
wip refactor ref image bufs
2025-10-02 16:37:17 +08:00
ac6be8ab8c
sd: do not force T5 on CPU anymore ( #1769 )
...
We now have the clip_cpu config parameter for that. Todo: Will make clip cpu on by default
2025-10-02 10:48:16 +08:00
4587ccb71a
prepare to refactor reference image
2025-10-02 10:41:29 +08:00
e49ac6b120
allow clip_vision to be loaded via clip_l or clip_g param
2025-10-01 17:57:49 +08:00
2fc31d36c0
gif mime type for animated images
2025-10-01 17:18:00 +08:00
9e4c29fda7
generate both gif and pick smaller (+1 squashed commits)
...
Squashed commits:
[09122d052] generate both gif and pick the smaller one
2025-09-30 14:42:58 +08:00
4117542eae
switch to msf gif
2025-09-30 13:56:41 +08:00
4b1c89ca5c
can save animated gifs
2025-09-29 22:52:42 +08:00
7b71742223
added additional toggles for SD, sdoffloadcpu , sdvaecpu and sdclipcpu
2025-09-28 22:08:29 +08:00
553890f975
compact UI launcher, WIP handling video outputs
2025-09-28 17:24:04 +08:00
cd6c771bd9
WAN single images working
2025-09-28 15:01:19 +08:00
cbb8a3feb5
taesd disable if not valid model arch
2025-09-27 17:00:30 +08:00
Concedo
Wagner Bruna
tsite