aboutsummaryrefslogtreecommitdiffstats
path: root/modules
AgeCommit message (Collapse)AuthorLines
2023-11-29Only apply ztSNR related code if alphas_cumprod existscatboxanon-9/+10
2023-11-29Lintcatboxanon-2/+2
2023-11-29Fix infotext for ztSNRcatboxanon-2/+2
2023-11-29Implement zero terminal SNR schedule optiondrhead-0/+28
2023-11-29Add options for zero terminal SNRdrhead-0/+2
2023-11-29Protect alphas_cumprod from downcastingdrhead-0/+5
2023-11-28Tweaked default values.CodeHatchling-4/+4
2023-11-28Tweaked UI sliders and labels.CodeHatchling-4/+4
2023-11-28Added slider for detail preservation strength, removed largely needless ↵CodeHatchling-20/+19
offset parameter, changed labels in UI and for saving to/pasting data from PNG files.
2023-11-28Added UI elements to control blending parameters.CodeHatchling-1/+62
2023-11-28Nerfs the aggressive post-processing step of overlaying the original image.CodeHatchling-3/+11
2023-11-28Implements "scheduling" for blending of the original latents and a latent ↵CodeHatchling-2/+59
blending formula that preserves details in blend transition areas.
2023-11-28Removed conflicting step that replaces the softly inpainted latents with a ↵CodeHatchling-3/+0
naive blend with the original latents.
2023-11-28Removed code which forces the inpainting mask to be 0 or 1. Now fractional ↵CodeHatchling-5/+1
values (e.g. 0.5) are accepted.
2023-11-29Fix the Ruff error about unused importhidenorly-1/+0
2023-11-29Add FP32 fallback support on torch.nn.functional.interpolatehidenorly-0/+16
This tries to execute interpolate with FP32 if it failed. Background is that on some environment such as Mx chip MacOS devices, we get error as follows: ``` "torch/nn/functional.py", line 3931, in interpolate return torch._C._nn.upsample_nearest2d(input, output_size, scale_factors) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ RuntimeError: "upsample_nearest2d_channels_last" not implemented for 'Half' ``` In this case, ```--no-half``` doesn't help to solve. Therefore this commits add the FP32 fallback execution to solve it. Note that the ```upsample_nearest2d``` is called from ```torch.nn.functional.interpolate```. And the fallback for torch.nn.functional.interpolate is necessary at ```modules/sd_vae_approx.py``` 's ```VAEApprox.forward``` ```repositories/stable-diffusion-stability-ai/ldm/modules/diffusionmodules/openaimodel.py``` 's ```Upsample.forward```
2023-11-29Revert "Add FP32 fallback support on sd_vae_approx"hidenorly-7/+1
This reverts commit 58c19545c83fa6925c9ce2216ee64964eb5129ce. Since the modification is expected to move to mac_specific.py (https://github.com/AUTOMATIC1111/stable-diffusion-webui/pull/14046#issuecomment-1826731532)
2023-11-28reformat file with uniform indentationw-e-w-104/+106
2023-11-28fix Auto focal point crop for opencv >= 4.8.xw-e-w-15/+18
autocrop.download_and_cache_models in opencv >= 4.8 the face detection model was updated download the base on opencv version returns the model path or raise exception
2023-11-27Update devices.pyobsol-1/+1
fixes issue where "--use-cpu" all properly makes SD run on CPU but leaves ControlNet (and other extensions, I presume) pointed at GPU, causing a crash in ControlNet caused by a mismatch between devices between SD and CN https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/14097
2023-11-27Allow use of mutiple styles csv filesCharlie Joynt-32/+171
* https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/14122 Fix edge case where style text has multiple {prompt} placeholders * https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/14005
2023-11-27catch uncaught exception with ui creation scriptsw-e-w-25/+29
prevent total webui crash
2023-11-27add Block component creation callbackw-e-w-0/+10
2023-11-26also consider extension urlJabasukuriputo Wang-1/+2
2023-11-26add categories to settingsAUTOMATIC1111-28/+96
2023-11-26json.dump(ensure_ascii=False)w-e-w-5/+5
improve json readability
2023-11-26compact prompt layout: preserve scroll when switching between lora tabsAUTOMATIC1111-1/+4
2023-11-26Merge branch 'dev' into test-fp8Kohaku-Blueleaf-69/+238
2023-11-26update torch to 2.1.0AUTOMATIC1111-5/+5
2023-11-26Merge pull request #14059 from akx/upruffAUTOMATIC1111-1/+1
Update Ruff to 0.1.6
2023-11-26do not save HTML explanations from options page to configAUTOMATIC1111-2/+2
2023-11-26Merge pull request #14084 from wfjsw/move-from-sysinfo-to-errorsAUTOMATIC1111-19/+17
Move exception_records related methods to errors.py
2023-11-26Merge branch 'hypertile-in-sample' into devAUTOMATIC1111-403/+13
2023-11-26Merge pull request #13948 from aria1th/hypertile-in-sampleAUTOMATIC1111-19/+404
support HyperTile optimization
2023-11-26rework hypertile into a built-in extensionAUTOMATIC1111-32/+13
2023-11-26move fileAUTOMATIC1111-371/+0
2023-11-25Fix pre-fp8Kohaku-Blueleaf-1/+1
2023-11-24remove traceback in sysinfoJabasukuriputo Wang-1/+0
2023-11-24Move exception_records related methods to errors.pyJabasukuriputo Wang-18/+17
2023-11-22use extension name for determining an extension is installed in the indexwfjsw-6/+8
2023-11-22Simplify restart_sampler (suggested by ruff)Aarni Koskela-1/+1
2023-11-22Add forced reload for fp16 cacheKohaku-Blueleaf-1/+1
2023-11-21Option for using fp16 weight when apply loraKohaku-Blueleaf-3/+13
2023-11-21fix [Bug]: (Dev Branch) Placing "Dimensions" first in "ui_reorder_list" ↵AUTOMATIC1111-6/+6
prevents start #14047
2023-11-21Add FP32 fallback support on sd_vae_approxhidenorly-1/+7
This tries to execute interpolate with FP32 if it failed. Background is that on some environment such as Mx chip MacOS devices, we get error as follows: ``` "torch/nn/functional.py", line 3931, in interpolate return torch._C._nn.upsample_nearest2d(input, output_size, scale_factors) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ RuntimeError: "upsample_nearest2d_channels_last" not implemented for 'Half' ``` In this case, ```--no-half``` doesn't help to solve. Therefore this commits add the FP32 fallback execution to solve it. Note that the submodule may require additional modifications. The following is the example modification on the other submodule. ```repositories/stable-diffusion-stability-ai/ldm/modules/diffusionmodules/openaimodel.py class Upsample(nn.Module): ..snip.. def forward(self, x): assert x.shape[1] == self.channels if self.dims == 3: x = F.interpolate( x, (x.shape[2], x.shape[3] * 2, x.shape[4] * 2), mode="nearest" ) else: try: x = F.interpolate(x, scale_factor=2, mode="nearest") except: x = F.interpolate(x.to(th.float32), scale_factor=2, mode="nearest").to(x.dtype) if self.use_conv: x = self.conv(x) return x ..snip.. ``` You can see the FP32 fallback execution as same as sd_vae_approx.py.
2023-11-20added option for default behavior of dir buttonsTom Haelbich-2/+8
2023-11-20Merge pull request #14009 from ↵AUTOMATIC1111-4/+21
AUTOMATIC1111/Option-to-show-batch-img2img-results-in-UI Option to show batch img2img results in UI
2023-11-20Merge branch 'dag' into devAUTOMATIC1111-153/+148
2023-11-20Merge pull request #13944 from wfjsw/dagAUTOMATIC1111-24/+188
implementing script metadata and DAG sorting mechanism
2023-11-20rework extensions metadata: use custom sorter that doesn't mess the order as ↵AUTOMATIC1111-153/+148
much and ignores cyclic errors, use classes with named fields instead of dictionaries, eliminate some duplicated code
generated by cgit v1.2.3 (git 2.39.1) at 2025-10-17 19:36:39 +0000