Is there an existing issue for this?
What happened?
When using the XYZ plot to generate images across multiple checkpoints, all but the first checkpoint will generate images differently than if you were to generate them without an XYZ plot.
For example, if you use "Model A" to generate one image without using XYZ plot, it will look like this:
If you use an XYZ plot to generate images from "Model A" followed by "Model B", the "Model A" image will look the same as above, as you would expect. However, if you set it as "Model B" followed by "Model A", then the image will come out differently, like this:
This happens for all but the first model in the list. This issue is not present on versions 1.2.1 or earlier, only 1.3.0.
Steps to reproduce the problem
Generate an image using any model ("Model A") using any settings without the XYZ plot.
Generate an image using the same settings, but while using an XYZ plot -> Checkpoint name. "Model A" must be the second checkpoint on the list, use any different model ("Model B") for the first checkpoint.
The two images from "Model A" will be different.
What should have happened?
The images from "Model A" should be the same. This was the behavior in previous versions.
Commit where the problem happens
What Python version are you running on ?
Python 3.10.x
What platforms do you use to access the UI ?
Windows
What device are you running WebUI on?
Nvidia GPUs (RTX 20 above)
What browsers do you use to access the UI ?
Mozilla Firefox
Command Line Arguments
@echo off
set COMMANDLINE_ARGS=--opt-channelslast --api --gradio-img2img-tool color-sketch --disable-nan-check --no-half-vae
call webui.bat
List of extensions
stable-diffusion-webui-pixelization
stable-diffusion-webui-state
Console logs
venv "C:\- Stable Diffusion\1.3.0\venv\Scripts\Python.exe"
fatal: not a git repository (or any of the parent directories): .git
fatal: not a git repository (or any of the parent directories): .git
Python 3.10.6 (tags/v3.10.6:9c7b4bd, Aug 1 2022, 21:53:49) [MSC v.1932 64 bit (AMD64)]
Version: <none>
Commit hash: <none>
Installing requirements
Launching Web UI with arguments: --opt-channelslast --api --gradio-img2img-tool color-sketch --disable-nan-check --no-half-vae
No module 'xformers'. Proceeding without it.
Loading weights [d8420de8d8] from C:\- Stable Diffusion\1.3.0\models\Stable-diffusion\Trained\ A-Zovya \ V3 Main \ aZovyaRPGArtistTools_v3.safetensors
Running on local URL: http://127.0.0.1:7860
To create a public link, set `share=True` in `launch()`.
Startup time: 5.2s (import torch: 1.3s, import gradio: 1.0s, import ldm: 0.5s, other imports: 0.7s, list SD models: 0.2s, load scripts: 0.8s, create ui: 0.3s, gradio launch: 0.2s).
Creating model from config: C:\- Stable Diffusion\1.3.0\configs\v1-inference.yaml
LatentDiffusion: Running in eps-prediction mode
DiffusionWrapper has 859.52 M params.
Applying optimization: sdp-no-mem... done.
Textual inversion embeddings loaded(0):
Model loaded in 5.0s (load weights from disk: 0.8s, create model: 0.6s, apply weights to model: 0.7s, apply channels_last: 0.6s, apply half(): 0.7s, move model to device: 0.6s, load textual inversion embeddings: 0.9s).
100%|██████████████████████████████████████████████████████████████████████████████████| 30/30 [00:06<00:00, 4.90it/s]
Total progress: 100%|██████████████████████████████████████████████████████████████████| 30/30 [00:05<00:00, 5.73it/s]
X/Y/Z plot will create 2 images on 1 2x1 grid. (Total steps to process: 60)████████████| 30/30 [00:05<00:00, 6.00it/s]
Loading weights [5baad911c8] from C:\- Stable Diffusion\1.3.0\models\Stable-diffusion\Trained\ A-Zovya \ V2-Offset \ aZovyaRPGArtistTools_v2Offset.safetensors
Applying optimization: sdp-no-mem... done.
Weights loaded in 1.4s (load weights from disk: 0.2s, apply weights to model: 0.7s, move model to device: 0.5s).
100%|██████████████████████████████████████████████████████████████████████████████████| 30/30 [00:05<00:00, 5.64it/s]
Loading weights [d8420de8d8] from C:\- Stable Diffusion\1.3.0\models\Stable-diffusion\Trained\ A-Zovya \ V3 Main \ aZovyaRPGArtistTools_v3.safetensors
Applying optimization: sdp-no-mem... done.
Weights loaded in 1.4s (load weights from disk: 0.2s, apply weights to model: 0.7s, move model to device: 0.5s).
100%|██████████████████████████████████████████████████████████████████████████████████| 30/30 [00:05<00:00, 5.67it/s]
Total progress: 100%|██████████████████████████████████████████████████████████████████| 60/60 [00:15<00:00, 3.79it/s]
Total progress: 100%|██████████████████████████████████████████████████████████████████| 60/60 [00:15<00:00, 6.10it/s]Additional information
No response
Is there an existing issue for this?
What happened?
When using the XYZ plot to generate images across multiple checkpoints, all but the first checkpoint will generate images differently than if you were to generate them without an XYZ plot.
For example, if you use "Model A" to generate one image without using XYZ plot, it will look like this:
If you use an XYZ plot to generate images from "Model A" followed by "Model B", the "Model A" image will look the same as above, as you would expect. However, if you set it as "Model B" followed by "Model A", then the image will come out differently, like this:
This happens for all but the first model in the list. This issue is not present on versions 1.2.1 or earlier, only 1.3.0.
Steps to reproduce the problem
Generate an image using any model ("Model A") using any settings without the XYZ plot.
Generate an image using the same settings, but while using an XYZ plot -> Checkpoint name. "Model A" must be the second checkpoint on the list, use any different model ("Model B") for the first checkpoint.
The two images from "Model A" will be different.
What should have happened?
The images from "Model A" should be the same. This was the behavior in previous versions.
Commit where the problem happens
What Python version are you running on ?
Python 3.10.x
What platforms do you use to access the UI ?
Windows
What device are you running WebUI on?
Nvidia GPUs (RTX 20 above)
What browsers do you use to access the UI ?
Mozilla Firefox
Command Line Arguments
@echo off set COMMANDLINE_ARGS=--opt-channelslast --api --gradio-img2img-tool color-sketch --disable-nan-check --no-half-vae call webui.batList of extensions
stable-diffusion-webui-pixelization
stable-diffusion-webui-state
Console logs
Additional information
No response