Stable-Diffusion-Regularization-Images
Stable-Diffusion-Regularization-Images | SD-Regularization-Images-Style-Dreambooth | |
---|---|---|
14 | 7 | |
99 | 29 | |
- | - | |
10.0 | 10.0 | |
over 1 year ago | over 1 year ago | |
- | - |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
Stable-Diffusion-Regularization-Images
-
Clarification regularization for Stable Diffusion
However, when I look at regularization dataset that people have created, a lot of them are composed by bad quality AI generated pictures, for instance disfigured humans, or images full or artifacts. For instance, this image of train, or this one of a woman.
- Training Picture Source
-
💡 How train with locally with 1.5 Runwayml Inpainting Model?
BTW, you can find the regularization images (ready to use class images) here.
-
Regularization images
Have you compared results to using regularization images from an existing repo such as https://github.com/JoePenna/Stable-Diffusion-Regularization-Images
-
Comic Diffusion V2. This is a culmination of everything worked towards so far. Trained on 6 styles at the same time, mix and match any number of them to create multiple different unique and consistent styles.
For subjects/people, paste this into the github downloader https://github.com/JoePenna/Stable-Diffusion-Regularization-Images/tree/main/person_ddim
-
Good Dreambooth Formula
If you are using person, man or woman as class, you don't need to generate the images as there are a some github repos that have a bunch of them already generated for you to use. Nitrosocke also shared some, check my initial post for the link.
- Custom Model Comparison 1.4 vs 1.5 (something broke)
-
What should I do when want better results for a person that was already trained in the sd v1.4 version? Train the model, Dreambooth, or textual inversion embeddings?
I did some experiments with dreambooth training. Overall better results were when I have used 1500 "person" class and about 50 training images. It is vital to have different background and different clothes otherwise it will "bake it" into your token (e.g. same sweater will influence all the rendering with color or pattern as it will be "part of your token"). Now I need to test textual inversion and see the difference.
-
Any advice on how to use the dream booth colab with automatic?
As for what kind of images to use I've tried actual photos of people and images generated with Stable Diffusion and I've had pretty good results with both. I also tried using exclusively pictures of the person I'm training for everything and even that worked pretty well. All I can really say is that it seems to pay off if you keep an eye on the framing of your images - if the majority of your reference images cut off the upper 10% of the head for example then your model will tend to also produce images that cut off the upper 10% of the head. Oh, and I haven't tried it myself but this Github repository apparently has a ton of images specifically for use in DreamBooth.
-
How are you achieving decent results in DreamBooth? My images look terrible!
I've made sure all my images are only me, and clean images. I have tried using the unsplash regularization images from https://github.com/JoePenna/Stable-Diffusion-Regularization-Images. I've tried generating my own images from SD itself. I've tried 1k, 2k, 3k, 4k steps. I've tried more images of myself and fewer. I've tried using "man", "person", "face" as the class. All of it results in absolute garbage. I get outputs that consistently look like I'm 80 years old or a different ethnicity. Or just wrong... so wrong.
SD-Regularization-Images-Style-Dreambooth
- Comic Diffusion V2. This is a culmination of everything worked towards so far. Trained on 6 styles at the same time, mix and match any number of them to create multiple different unique and consistent styles.
-
Question about training styles
I'm using the Joe Penna's repo on runpod and using only 20 training images and 1700 reg images from https://github.com/aitrepreneur/SD-Regularization-Images-Style-Dreambooth to trin styles. I'm getting very good results.
-
Classic Disney animation dreambooth model
I'm new to using dreambooth, but I followed the steps in some of the recent trending examples to make a "classic disney" art style. I pulled/cropped/reframed about 50 reference images, and used the style examples [from here](https://github.com/aitrepreneur/SD-Regularization-Images-Style-Dreambooth), trained with 6400 steps. Colors are typically oversaturated, and it's really hard to control. I've also found that adding artists helps balance the composition out a lot. Here are some of the sample outputs!
-
Fine-tuned the model on Kurzgesagt videos with DreamBooth. Here are some results.
I've used this repository for regularization images. And these options for training: --class_word "style" --token "kurzgesagt"
- 2D Illustration Styles are scarce on Stable Diffusion so i created a dreambooth model inspired by Hollie Mengert's work
- Hello, i saw that you can train dreambooth for a style, I tried taring dreambooth on vast.ai for a children book illustration style but I got pretty awful result, any ideas what went wrong.
-
I've further refined my Studio Ghilbi Model
I used around 20,000 steps (I forgot to look at number of steps when I stopped training). The regulation images I used can be obtained at https://github.com/aitrepreneur/SD-Regularization-Images-Style-Dreambooth
What are some alternatives?
Dreambooth-Stable-Diffusion - Implementation of Dreambooth (https://arxiv.org/abs/2208.12242) by way of Textual Inversion (https://arxiv.org/abs/2208.01618) for Stable Diffusion (https://arxiv.org/abs/2112.10752). Tweaks focused on training faces, objects, and styles.
Dreambooth-SD-optimized - Implementation of Dreambooth (https://arxiv.org/abs/2208.12242) with Stable Diffusion
Dreambooth-Regularization - All the regs
stable-diffusion-webui - Stable Diffusion web UI
Txt2Vectorgraphics - Custom Script for Automatics1111 StableDiffusion-WebUI.
Dreambooth-Stable-Diffusion - Implementation of Dreambooth (https://arxiv.org/abs/2208.12242) with Stable Diffusion (tweaks focused on training faces)
diffusers - 🤗 Diffusers: State-of-the-art diffusion models for image and audio generation in PyTorch
InvokeAI - InvokeAI is a leading creative engine for Stable Diffusion models, empowering professionals, artists, and enthusiasts to generate and create visual media using the latest AI-driven technologies. The solution offers an industry leading WebUI, supports terminal use through a CLI, and serves as the foundation for multiple commercial products.