Being able to create our favorite characters is liberating, isn't it? For people like me who also really enjoy generating content with AI, however, there are some limiting factors, such as knowing whether or not the checkpoint we are using is trained for that particular character. If it isn't, we need to use a resource called LoRA. Of course, there are many other functions for a LoRA, and I'm just scratching the surface here.
One of my favorite characters is Rook, a wolf who is the protagonist of the doujinshi "Happening at School," created by Canyne Khai. I've been visiting CivitAI for months and looking for open-source models for this character, but without success. What I managed to find, as some of you have already noticed, is "Brown Dog," who makes a special appearance in this same doujinshi.
So, after Charliemon taught me how to make LoRAs for characters, I decided to take the initiative and do it myself, and you can see the result above. The training base is somewhat limited and of average quality, because unlike Dainen, where I can generate as many high-quality images as I need, here I had to use what I could find.
Anyway, from now on we can see this sexy boy not only in my gallery, because at the time of writing this post, this LoRA is publicly available for both Pony and Illustrious models and can be downloaded for free at this link: https://l.dainendog.com/brad-lora. And even better: if you don't have a good computer or don't know how to use Stable Diffusion locally, you can use CivitAI's own image creator for free, since my LoRA has been approved for being used there as well.
Of course, everything I publish here follows the site's rules, and all content will be available for free on my CivitAI profile, which you can follow by opening this link: https://l.dainendog.com/civitai.
Thank you once again for the support and help you have given me since the beginning, and once again a special thanks to Charliemon for helping me take the first steps on this new and exciting journey!
Notes: - The images were not corrected on purpose, to show the result without any type of improvement or corrections. - This is an extra post and will not affect my third post of the day, which should be online tonight, so stay tuned!
