mirror of
https://github.com/easydiffusion/easydiffusion.git
synced 2024-11-28 03:03:11 +01:00
Updated Custom Models (markdown)
parent
67ae7df6b3
commit
7154cfa2cc
@ -1,11 +1,11 @@
|
||||
## What are Stable Diffusion models?
|
||||
Stable Diffusion UI uses so called models to create the images. The models get trained using many images and image descriptions. During the installation,
|
||||
a default model gets downloaded, the sd-v1-4 model. Other models exist. Some of them use sd-v1-4 as their base and were then trained on additional images, while other models were trained from scratch.
|
||||
Easy Diffusion uses "models" to create the images. These models get trained using many images and image descriptions. During the installation,
|
||||
a default model gets downloaded, the sd-v1-5 model. Other models exist. Some of them use sd-v1-5 as their base and were then trained on additional images, while other models were trained from scratch.
|
||||
|
||||
Depending on the images used during the training, the models generate a different look and understand different prompts.
|
||||
|
||||
## Installing a model
|
||||
If you've downloaded a model, you can make it available to Stable Diffusion UI:
|
||||
If you've downloaded a model, you can make it available to Easy Diffusion:
|
||||
- Copy the model file to the `models\stable-diffusion` folder (inside the installed folder). The file should have the suffix `.ckpt` or `.safetensors`.
|
||||
- If the model comes with a `.yaml` file, please place that yaml file next to the model, with the same name. For e.g. if the model is `some-model.ckpt` (or `some-model.safetensors`), then please rename the yaml file to `some-model.yaml`, and place it next to the model file in the same folder.
|
||||
- Refresh the model list using the "reload"-icon in the "Models" section in the image settings:
|
||||
@ -14,15 +14,6 @@ If you've downloaded a model, you can make it available to Stable Diffusion UI:
|
||||
|
||||
The model should now be visible in the "Models" dropdown.
|
||||
|
||||
### Troubleshooting - Black images?
|
||||
Please add this to the **end** of the `.yaml` file that came with your model:
|
||||
```yaml
|
||||
extra:
|
||||
attn_precision: "fp32"
|
||||
```
|
||||
|
||||
For e.g. please see the end of this file for reference: https://github.com/easydiffusion/sdkit/blob/main/sdkit/models/models_db/configs/v2.1-inference-v.yaml#L69
|
||||
|
||||
## Sharing model files with other UIs on your PC
|
||||
If you use multiple Stable Diffusion UIs, you can save disk space by keeping a common copy of the model files. You can create a linked folder to share the models folder between the different Stable Diffusion software.
|
||||
|
||||
|
Loading…
Reference in New Issue
Block a user