License CC BY-SA Protocol (Free modification, but with attribution to the original protocol author)
<aside> ✨ This is a Beta version, and with the help of the community, it will be optimized further in future versions. It focuses on the creation of LORAs via Colab (as this is the process being streamlined). I that someone with a good LORA creation team could make them even more quickly.
</aside>
This is not a step-by-step guide, but a way to focus on creating LORAs to save time and train many in a short period. I will not provide specific information on how to create the dataset, tag, or train, but will provide general advice. Technical information can be found in this guide: https://civitai.com/models/22530/guide-make-your-own-loras-easy-and-free
Of course, you don't have to be so organized. You can create your LORAs in your own way, but anyone who enjoys organizing and simplifying processes will love this guide. Also, those who follow or improve the speed of LORA creation will inevitably stand out in Civitai due to the quick creation of characters.
<aside> <img src="/icons/send-to_blue.svg" alt="/icons/send-to_blue.svg" width="40px" /> It is recommended that you use the following Notion template to facilitate the LORA creation process. This is the template I use and will be mentioned in various parts of this guide:
LORA project manager via Notion
</aside>
This protocol will allow:
Anyone to have everything more organized.
Us to scale model training by creating work groups with several people to perform LORA in an organized and much faster way).
This means, for example, that some people will be responsible for creating the dataset and others for the training, or some for testing the results, in an organized group.
Making up to 12 quality LORAs per week without wearing ourselves out (at least individually).
<aside> 👉 It's a good idea to keep a list to keep track of the LORAs. You can use a "To-Do" app like "Microsoft To-Do" or the template I'll provide later in Notion.
</aside>
