[Bug]: v-pred checkpoint seems unable to be loaded correctly

[whitesky1226]

Checklist

• The issue exists after disabling all extensions
• The issue exists on a clean installation of webui
• The issue is caused by an extension, but I believe it is caused by a bug in the webui
• The issue exists in the current version of the webui
• The issue has not been reported before recently
• The issue has been reported before but has not been fixed yet

What happened?

The checkpoint trained by v-pred method seems unable to be loaded correctly. In the original A1111 webui, the ckpt model can be loaded correctly and generates images normally by placing a yaml configuration file with the same name as the ckpt in the ckpt folder. However, this does not seem to work in the current Forge version. After switching to the vpred checkpoint, it only generates noise.

Steps to reproduce the problem

1. Switching to a vpred checkpoint model.
2. Txt2img as normal.
3. A picture full of noise is generated.

What should have happened?

The yaml file of the ckpt model is not loaded.

What browsers do you use to access the UI ?

Google Chrome

Sysinfo

none

Console logs

the original A1111 webui 1.7.0:

Loading model yiffymix_v41.safetensors [b8b4c24fb6] (2 out of 2)
Loading weights [b8b4c24fb6] from F:\AI-painting\AUTOMATIC1111-sd-webui\models\Stable-diffusion\yiffymix_v41.safetensors
Creating model from config: F:\AI-painting\AUTOMATIC1111-sd-webui\models\Stable-diffusion\yiffymix_v41.yaml
Loading VAE weights specified in settings: F:\AI-painting\AUTOMATIC1111-sd-webui\models\VAE\vae-ft-mse-840000-ema-pruned.safetensors
Applying attention optimization: sdp... done.
Model loaded in 4.1s (load weights from disk: 1.5s, create model: 0.3s, apply weights to model: 0.7s, load VAE: 0.3s, load textual inversion embeddings: 1.1s).

***************************************************************
the Forge version:

Loading weights [b8b4c24fb6] from F:\AI-painting\stable-diffusion-webui-forge\models\Stable-diffusion\yiffymix_v41.safetensors
model_type EPS
UNet ADM Dimension 0
Using pytorch attention in VAE
Working with z of shape (1, 4, 32, 32) = 4096 dimensions.
Using pytorch attention in VAE
Loading VAE weights specified in settings: F:\AI-painting\stable-diffusion-webui-forge\models\VAE\vae-ft-mse-840000-ema-pruned.safetensors
To load target model SD1ClipModel
Begin to load 1 model
Model loaded in 4.9s (unload existing model: 1.0s, load weights from disk: 1.3s, forge load real models: 1.3s, load VAE: 0.3s, load textual inversion embeddings: 0.9s).

Additional information

No response

[V1Htor]

Same issue. Vpred models produce white images.

[lllyasviel]

give me an example model that i can test

[V1Htor]

give me an example model that i can test

I use model + yaml + lyco from here https://rentry.org/5exa3

[lllyasviel]

update and try again

[whitesky1226]

Fixed. 谢谢😊

[V1Htor]

update and try again

amazing speed!