This model is just the first dreambooth iteration for concept-sheet/old-books style. Based on SD 2.1 Installation: As it is model based on 2.1 to make it work you need to use .yaml file with name of a model (vector-art.yaml). The yaml file is included here as well to download. Simply copy paste to the same folder as selected model file. Usually this is the <code>models/Stable-diffusion</code> one.License and support:If you like what I am doing and you want to support my current and upcoming projects: here is my <a target="_blank" rel="ugc" href="https://www.buymeacoffee.com/pmejna">coffee jar</a>You can use this model as you like. I love open source so have fun with it. :) If you like it I appreciate your support ^^ Cheers!Black images issue:2.1 models need to have a web-ui config modified - if you are getting black images - go to your config file and add to COMMANDLINE_ARGS= <code>--no-half</code> - potentially it could work with <code>--xformers</code> instead (if supported). This line might slow your generations a bit but will not affect negatively your output.<img src="https://imagecache.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/e717898c-c997-4f10-e832-04e1b0e1c000/width=525" /> You will find example props on the images. Embeddings:This model works really well with various embeddings. I would recommend you especially these two:<ul><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/4775/planit-a-documentation-embedding">https://civarchive.com/models/4775/planit-a-documentation-embedding</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/4805/protogemb-2-for-sd-2x-768">https://civarchive.com/models/4805/protogemb-2-for-sd-2x-768</a></li></ul>

concept-sheet

This version is doing what is intended but is still far from what I want to achieve. You can try using the keyword <code>concept-art</code> to strengthen the effect. In the future version, I would like to add more variety to a character concept. Currently, this seems overfitted. Waiting for dream booth collab fix supporting session to resume. I will be probably training on a few hundred selected images.

Files with SHA256: 5926432c85dc32d744a36b10de2918713436ba1812857d0d7faf169a817149f4