Loose and free oil painting
The look is of a study, or the initial blocking in before detailed rendering, or just the work of a painter who eschews fine details.
Colors lean toward naturalistic earth tones, but you can push it to be a bit more colorful in your prompt, to an extent.
The output of this model, trained on the same dataset used for its SDXL version, was initially much more rough, dark, and brown than I care for. So I merged it with my Eldritch Oil painting model and an unreleased oil painting model that I use just for such merges. This brings a pleasing color balance and the crucial features of a scene, such as faces and hands, will tend to be rendered in more detail, while the environment often receives very gestural, abstract treatment.
Description
initial version
FAQ
Comments (11)
love love love this lora
thanks @barepixels!
Like this thing a lot. That's pretty much how a attempt everything using traditional/digital media. Gesture, colors, values, rough painting etc... even the brush kinda matches.
Definitely one of the best Flux LoRAs I've used. Flexible, coherent, blends well with other LoRAs and the style is spot on.
Thank you again for sharing your creations 🙏👍
really appreciate the kind words, but even more so I always appreciate seeing what imagery you create!
@eldritchadam Thank you, it is always a pleasure to get encouragements from you 🙏
Fab lora, so good.
If you like this LoRA, you'll love @robadams171's OilPaintz LoRA - check it out! Very similar style, but better color range and more consistent, painterly treatment of faces. Really excellent!
A huge thank you to Eldritch for creating the amazing Eldritch Sketch Painting for Flux.1 D LORA model! I absolutely love it, and I’m pretty sure the over 2,800 likes I’ve received for one image are mostly thanks to this incredible LORA. My workflow might have helped a little, but the real star here is your creation. Keep up the fantastic work! 🎨✨
Definetly the best Lora for Flux, I use it litteraly everytime ! It is coherant, blends perfectly with most other loras and is overall gorgeous! Thank you @eldritchadam !
Now I just wished there was a version for Qwen too or the files it has been trained on available somewhere so I could trained a Lora for Qwen myself.
thanks for the kind words! If you want the dataset, you can have it at this Google Drive link:
https://drive.google.com/drive/folders/1gWJrlbZxmkfGhUFF-6rBUWyWRl-a8MnL?usp=drive_link
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.