Is there an existing issue for this?
What happened?
After last update [9aba26abdf] X/Y/Z plot load different model than picked in settings.
This is generation from X/Y/Z plot
Those from normal txt2img generated one by one
deliberate_v2
hands_new2_with_multinoise
hands_new2_with_noise
As you can see image from hands_new2_with_noise is different than others when generated separatly but it's exactly the same as hands_new2_with_multinoise when generated with X/Y/Z plot
Steps to reproduce the problem
- Go to txt2img
- Choose X/Y/Z plot
- Pick models with similar names (same prefix?)
What should have happened?
Each image should be different
Commit where the problem happens
9aba26abdf
What platforms do you use to access the UI ?
Windows
What browsers do you use to access the UI ?
Google Chrome
Command Line Arguments
List of extensions
SD-CN-Animation
clip-interrogator-ext
ddetailer
loopback_scaler
multidiffusion-upscaler-for-automatic1111
sd-dynamic-thresholding
sd-infinity-grid-generator-script
sd-webui-controlnet
sd-webui-cutoff
sd-webui-regional-prompter
sd-webui-supermerger
sd-webui-text2video
sd-webui-tome
sd_webui_SAG
seed_travel
LDSR
Lora
ScuNET
SwinIR
prompt-bracket-checker
Console logs
X/Y/Z plot will create 3 images on 1 3x1 grid. (Total steps to process: 60)████████████| 40/40 [00:40<00:00, 1.31it/s]
100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [00:16<00:00, 1.24it/s]
Loading weights [9aba26abdf] from E:\stable-diffusion-webui\models\Stable-diffusion\deliberate_v2.safetensors 1.30it/s]
Loading VAE weights specified in settings: E:\stable-diffusion-webui\models\VAE\vae-ft-mse-840000-ema-pruned.ckpt
Applying cross attention optimization (Doggettx).
Weights loaded in 2.3s (load weights from disk: 0.2s, apply weights to model: 0.5s, load VAE: 0.3s, move model to device100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [00:16<00:00, 1.24it/s]
Loading weights [7d13449163] from E:\stable-diffusion-webui\models\Stable-diffusion\hands_new2_with_multinoise.safetensors
Loading VAE weights specified in settings: E:\stable-diffusion-webui\models\VAE\vae-ft-mse-840000-ema-pruned.ckpt
Applying cross attention optimization (Doggettx).
Weights loaded in 2.7s (load weights from disk: 0.6s, apply weights to model: 0.5s, load VAE: 0.3s, move model to device: 1.4s).
100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [00:16<00:00, 1.25it/s]
Total progress: 100%|██████████████████████████████████████████████████████████████████| 60/60 [00:58<00:00, 1.03it/s]
Additional information
No response
Is there an existing issue for this?
What happened?
After last update [9aba26abdf] X/Y/Z plot load different model than picked in settings.
This is generation from X/Y/Z plot
Those from normal txt2img generated one by one
deliberate_v2
hands_new2_with_multinoise
hands_new2_with_noise
As you can see image from hands_new2_with_noise is different than others when generated separatly but it's exactly the same as hands_new2_with_multinoise when generated with X/Y/Z plot
Steps to reproduce the problem
What should have happened?
Each image should be different
Commit where the problem happens
9aba26abdf
What platforms do you use to access the UI ?
Windows
What browsers do you use to access the UI ?
Google Chrome
Command Line Arguments
List of extensions
SD-CN-Animation
clip-interrogator-ext
ddetailer
loopback_scaler
multidiffusion-upscaler-for-automatic1111
sd-dynamic-thresholding
sd-infinity-grid-generator-script
sd-webui-controlnet
sd-webui-cutoff
sd-webui-regional-prompter
sd-webui-supermerger
sd-webui-text2video
sd-webui-tome
sd_webui_SAG
seed_travel
LDSR
Lora
ScuNET
SwinIR
prompt-bracket-checker
Console logs
Additional information
No response