aboutsummaryrefslogtreecommitdiffstats
path: root/modules/sd_samplers_common.py
Commit message (Collapse)AuthorAgeFilesLines
* Add Pad conds v0 optionAUTOMATIC11112024-01-271-0/+7
|
* fix an issue where VAE would remain in fp16 after an auto-switch to fp32AUTOMATIC11112023-08-301-0/+2
|
* add an option to choose how to combine hires fix and refinerAUTOMATIC11112023-08-301-2/+11
|
* actual solution to the uncommon hanging problem that is seemingly caused by ↵AUTOMATIC11112023-08-221-1/+1
| | | | multiple progress requests working on same tensor
* attemped solution to the uncommon hanging problem that is seemingly caused ↵AUTOMATIC11112023-08-221-1/+1
| | | | by live previews working on the tensor as denoising
* forbid Full live preview method for medvram and add a setting to undo the ↵AUTOMATIC11112023-08-211-1/+6
| | | | forbidding
* Merge pull request #12630 from catboxanon/fix/nans-mk2AUTOMATIC11112023-08-191-1/+2
|\ | | | | Attempt to resolve NaN issue with unstable VAEs in fp32 mk2
| * Attempt to resolve NaN issue with unstable VAEs in fp32 mk2catboxanon2023-08-171-1/+2
| |
* | Add option for faster live interruptcatboxanon2023-08-181-1/+1
|/
* Remove wrong scaleKohaku-Blueleaf2023-08-161-2/+1
|
* correctly add Eta DDIM to infotext when it's 1.0 and do not add it when it's ↵AUTOMATIC11112023-08-141-1/+2
| | | | 0.0.
* Merge pull request #12514 from catboxanon/feat/batch-encodeAUTOMATIC11112023-08-131-1/+9
|\ | | | | Encode batch items individually to significantly reduce VRAM
| * Encode batches separatelycatboxanon2023-08-131-1/+9
| | | | | | | | | | Significantly reduces VRAM. This makes encoding more inline with how decoding currently functions.
* | Add s_noise param to more samplerscatboxanon2023-08-131-4/+4
|/
* use dataclass for StableDiffusionProcessingAUTOMATIC11112023-08-131-1/+4
|
* add refiner to StableDiffusionProcessing classAUTOMATIC11112023-08-131-1/+1
| | | | write out correct model name in infotext, rather than the refiner model
* make it possible to use hires fix together with refinerAUTOMATIC11112023-08-121-0/+3
|
* put refiner into main UI, into the new accordions sectionAUTOMATIC11112023-08-121-17/+23
| | | | | | add VAE from main model into infotext, not from refiner model option to make scripts UI without gr.Group fix inconsistencies with refiner when usings samplers that do more denoising than steps
* Merge branch 'dev' into refinerAUTOMATIC11112023-08-101-12/+12
|\
| * rework RNG to use generators instead of generating noises beforehandAUTOMATIC11112023-08-091-12/+12
| |
* | add None refiner optionAUTOMATIC11112023-08-081-0/+3
| |
* | merge errorsAUTOMATIC11112023-08-081-1/+5
| |
* | Merge branch 'dev' into refinerAUTOMATIC11112023-08-081-1/+137
|\|
| * add CFG denoiser implementation for DDIM, PLMS and UniPC (this is the commit ↵AUTOMATIC11112023-08-081-1/+139
| | | | | | | | when you can run both old and new implementations to compare them)
* | add infotextAUTOMATIC11112023-08-061-9/+20
| | | | | | | | | | add proper support for recalculating conds in k-diffusion samplers remove support for compvis samplers
* | initial refiner supportAUTOMATIC11112023-08-061-1/+18
|/
* repair broken live previews if using VAE with halfAUTOMATIC11112023-08-061-1/+1
|
* remove noneed scaleKohaku-Blueleaf2023-08-041-1/+1
|
* Revert "change all encode"Kohaku-Blueleaf2023-08-041-1/+1
| | | | This reverts commit 094c416a801b16c7d8e1944e2e9fae2c9e98bf12.
* change all encodeKohaku-Blueleaf2023-08-041-1/+1
|
* move the modified decode into smapler_commonKohaku-Blueleaf2023-08-041-3/+3
|
* Merge branch 'dev' into efficient-vae-methodsKohaku-Blueleaf2023-08-041-0/+6
|\
| * resolve some of circular import issues for kohakuAUTOMATIC11112023-08-041-2/+8
| |
* | Fix code styleKohaku-Blueleaf2023-08-041-2/+2
| |
* | add TAESD for i2i and t2iKohaku-Blueleaf2023-08-041-5/+33
|/
* rework torchsde._brownian.brownian_interval replacement to use ↵AUTOMATIC11112023-08-031-6/+6
| | | | device.randn_local and respect the NV setting.
* simplify single_sample_to_imageAUTOMATIC2023-05-171-6/+3
|
* TAESD fixSakura-Luna2023-05-171-4/+5
|
* Merge branch 'dev' into taesd-aAUTOMATIC11112023-05-171-1/+20
|\
| * suppress ENSD infotext for samplers that don't use itAUTOMATIC2023-05-161-1/+20
| |
* | return live preview defaults to how they wereAUTOMATIC2023-05-171-14/+15
| | | | | | | | | | | | only download TAESD model when it's needed return calculations in single_sample_to_image to just if/elif/elif blocks keep taesd model in its own directory
* | Add live preview mode checkSakura-Luna2023-05-141-2/+2
| |
* | Add Tiny AE live previewSakura-Luna2023-05-141-8/+13
|/
* rename CPU RNG to RNG source in settings, add infotext and parameters ↵AUTOMATIC2023-04-291-1/+2
| | | | copypaste support to RNG source
* Option to use CPU for random number generation.Deciare2023-04-191-0/+9
| | | | | | | Makes a given manual seed generate the same images across different platforms, independently of the GPU architecture in use. Fixes #9613.
* Refactor Mac specific code to a separate filebrkirch2023-02-011-16/+0
| | | | Move most Mac related code to a separate file, don't even load it unless web UI is run under macOS.
* split samplers into one more files for k-diffusionAUTOMATIC2023-01-301-1/+2
|
* split compvis sampler and shared sampler stuff into their own filesAUTOMATIC2023-01-301-477/+2
|
* Split history sd_samplers.py to sd_samplers_common.pyAndrey2023-01-301-0/+552
generated by cgit v1.2.3 (git 2.25.1) at 2025-04-28 20:58:08 +0000