title: boring_e621
tags:
- textual inversion embeddings
- image-generation
license: apache-2.0
boring_e621
This embedding attempts to capture what it means for an image to be uninteresting. It was trained as a negative embedding using e621 style tags as prompts during training.
If you're using the Automatic1111 Stable Diffusion WebUI, place the .pt file in
stable-diffusion-webui\embeddings and add "by boring_e621" to your negative prompt for more interesting outputs.
Model Description
The motivation for boring_e621 is that negative embeddings like Bad Prompt, whose training is described here depend on manually curated lists of tags describing features people do not want their images to have, such as "deformed hands". Some problems with this approach are:
- Manually compiled lists will inevitably be incomplete.
- Models might not always understand the tags well due to a dearth of training images labeled with these tags.
- It can only capture named concepts. If there exist unnamed yet visually unappealing concepts that just make an image look wrong, but for reasons that cannot be succinctly explained, they will not be captured by a list of tags.
To address these problems, boring_e621 employs textual inversion directly on a set of 100 of the most boring images on the furry image site e621.net.
These images were automatically collected by querying e621 for images that no user has ever Favorited or Up-Voted.
Almost all images on the site have at least one up vote or favorite, so the images that no one favorited are some of the very most uninspiring images on the website.
Thus, you should be able to use the resulting embedding as a negative prompt to make your generated images Not Boring.
Bias, Risks, and Limitations
- Using this as a negative embedding often sacrifices some fidelity to the prompt. For example, characters in the image may disappear or change eye/skin/fur color.
- Using this as a negative embedding may introduce unexpected or undesired content into the image to make it look less boring.
- Unlike other negative embeddings, this is not intended to fix problems like extra limbs or deformed hands. It can be used alongside other negative embeddings to fix deformities.
Evaluation
I extracted the tags from three e621 images and used them to construct a set of test prompts.
- one prompt was constructed from an image with a high number of favorites.
- one prompt was constructed from an image with a moderate number of favorites.
- one prompt was constructed from an image with 0 favorites.
I then generated test images from each of these prompts, each time using a different negative embedding as the negative prompt. Particularly, I tried:
Finally, I qualitatively evaluated the attractiveness and interestingness of the resulting images, though I will let you draw your own conclusions from the output below.
Results
Other Models
Boring_e621 has been reported to work well with SD 1.4 or 1.5 furry models such as: