---
tags:
- text-to-image
- lora
- diffusers
- template:diffusion-lora
widget:
- output:
    url: samples/orange_splatter.png
  text: 'Orange Splatter Illustration samples'
- output:
    url: samples/tldr_samples.png
  text: 'Teal Dark (TLDR) samples'
- output:
    url: samples/wireframe_samples.png
  text: 'Blueprint / Wireframe samples'
- output:
    url: samples/sts2_style.png
  text: 'Slay The Spire 2 samples'
- output:
    url: samples/cyanide_and_happiness.png
  text: 'Cyanide and Happiness samples'
- output:
    url: samples/arcane.png
  text: 'Arcane samples'
- output:
    url: samples/devil_may_cry.png
  text: 'Devil May Cry samples'
- output:
    url: samples/clothes_line.png
  text: 'Clothes line samples'
base_model: black-forest-labs/FLUX.2-klein-9B
instance_prompt: null
license: apache-2.0
---

# Flux.2 Klein 9b LORAs

To be used with the Flux.2 Klein 9b Distilled model.

<Gallery />

#### Orange Splatter Illustration Aesthetic

Use `dvr_osi_style` as the trigger word. Similar to Teal Dark, but Orange.


#### Teal Dark (TLDR) Aesthetic

Use `dvr_tldr_style` as the trigger word. Works with text to image (you can increase the strength is the style doesn't come through for more realistic prompts) and image to image workflows. For I2I add "black background" for a nice effect. [More examples in this Reddit post](https://www.reddit.com/r/StableDiffusion/comments/1t8bn7f/teal_dark_flux2_klein_9b_styleaesthetic_lora/)

#### Blueprint / Wireframe style

Use `dvr_wf_style` as the trigger word. Works with text to image and image to image workflows. [More examples in this Reddit post](https://www.reddit.com/r/StableDiffusion/comments/1t4udu2/wireframe_flux2_klein_9b_style_lora/)


#### Slay The Spire 2 (events) style

Use `sts2_style` as the trigger word. For best results recommended to use with additional modifiers like "dark fantasy illustration". Trained on 55 game events artwork.
[More examples in this Reddit post](https://www.reddit.com/r/StableDiffusion/comments/1sjahk9/slay_the_spire_2_flux2_klein_9b_style_loras/)

#### Slay The Spire 2 (events + 600 card artwork) style (experimental)

Use `sts2_style` as the trigger word. Trained on 55 events + 600 additional card artwork.
Personally I prefer the simple style but this one has been additionally captioned with each class style and each card name (examples `sts2_style, ironclad_card_style, card_aggression` ... , `sts2_style, regent_card_style, card_bombardment`, `sts2_style, card_alchemize`)

#### Cyanide and Happiness style (experimental)

Use `ch_visual_style, stick figure character` as the trigger word. Optional add more keywords to guide the style: "flat vector art, minimalist lineart". Can work as an edit lora too.

#### Arcane style

Use `arcane_visual_style` as the trigger word. The lora works for text to image but also as an edit lora (example: `Transform into arcane_visual_style` feel free to add extra words like "keep the pose" etc if you find the image drifts too much from the original). 
[More examples in this Reddit post](https://www.reddit.com/r/StableDiffusion/comments/1qm0tt6/arcane_flux2_klein_9b_style_lora_t2i_and_edit/)


#### Devil May Cry style

Use `dmc_style` as the trigger word. The lora works for text to image but also as an edit lora (example: `Transform into dmc_style` feel free to add extra words like "keep the pose" etc if you find the image drifts too much from the original). [samples](samples/devil_may_cry.png)

Known limitations: sometimes Flux likes to make photo realistic images when using certain keywords. For best results avoid references to camera names (example: DO NOT use `shot on Fuji Pro 400H` or similar).


#### Clothes line concept (experimental)

Use `clthLn, a ... made of clothes with pegs on a line, ...`. 
Dataset found from this [Reddit post](https://www.reddit.com/r/oddlysatisfying/comments/1s5awwa/photographer_creates_art_using_clothes_on_a/) and is mostly animals


#### Disclaimer and Important notes
The LoRAs in this repository are created for research purposes only and are not intended for commercial use. Some might be based on TV series which are protected by copyright. Users utilize these models at their own risk. Users are obligated to comply with copyright laws and applicable regulations. The models have been developed for non-commercial purposes, and it is not my intention to infringe on any copyright. I assume no responsibility for any damages or legal consequences arising from the use of these models.