Discovered this wonderful Parisian artist by the name Araminta K who has been exploring the latent space and embracing the possibilities. Araminta has trained several LoRAs that have been released on Huggingface page
Download the LoRA individually to run locally on your computer or explore and play with them online via Araminta’s Portfolio page.
At the time of publishing this post there were 11 13 LoRAs available that range from soft tones, water color illustration, pop art, sketch art and anime manga style. These LoRAs are designed for use with standard base SDXL model.
- Araminta K Illustration
- Japanese Poster Prints
- Abstract Anime Painting
- Gemini Anime
- Pop Art Anime
- Soft Ones
- Soft Pastel Anime
- Painting Light
- V. Sketch Manga
- Lost and Lonely Anime
- Sketched Out Manga
- Haute Couture Illustration (added 11th June)
- B&W Manga Block (added 11th June)
Exploring the different LoRAs the best way to do this is to lock the Seed and use the same prompt, the only thing that changes is the LoRA. You end up with a collection of images that you can compare easily to get a feel for how the result is influenced by the LoRA.
Araminta uses the keyword/trigger word “daiton style” to invoke the LoRA however I found that even without the trigger word the generation is influenced heavily by the LoRA especially if you are using the Base SDXL model
prompt: happy, a cute girl with black strange hair, red sweater, staring into the camera
seed: 3474888480
model: Base SDXL
Above grid of images showcases some of the aesthetics of the images you can expect to see however they can also be highly influenced by the prompt you use. The last image in the grid is the base SDXL render without any of the LoRA.
I used a specifically designed workflow in ComfyUI that allowed me to test these LoRA which I am going to release tomorrow so you can download it and have fun testing these and other LoRAs.
Of course to get really creative you can combine one or more LoRAs in your workflow and adjust the strength of each to blend the aesthetics together.
Here is a gallery of the images that was able to create using the Base SDXL model and the LoRAs shared by Araminta. Make sure you follow us for more updates and also the artist Araminta.
If you'd like to support our site please consider buying us a Ko-fi, grab a product or subscribe. Need a faster GPU, get access to fastest GPUs for less than $1 per hour with RunPod.io