Ev Ganin Style (Krea2 Style LoRA) - V1

If you love the Midjourney aesthetic, you have likely seen this distinct style. Ev Ganin is well-known for generating incredibly striking, bold line art combined with highly stylized, exaggerated cartoony proportions. I wanted to take that highly restricted, closed-ecosystem look and bring it natively to Krea 2 so the open-source community could have full, uninhibited control over it.

⚠️ V1 Usage, Strengths, & Limitations

I am releasing this V1 so you can start experimenting, but please be aware of a specific quirk! Because the dataset happened to be heavily female skewed, the model highly favors female subjects at higher strengths.

• Recommended Strength: Works best between 1.0 and 3.0.

• The Quirks (Gender Bias): Below 2.0, you get a great "inspired by" feel and the model obeys gender tags reasonably well. However, if you push the strength above 2.0, the style becomes much more pronounced, but the model will begin ignoring male/other prompts and heavily force a female character.

• The Plan for V2: I will be doing a V2 retrain soon with a heavily balanced dataset to fix the gender bleeding. For now, V1 is fun and highly stylized, just keep an eye on your weights!

Trigger Word: EvGaninStyle

🛠️ Training Details & Community Shoutout

• Dataset: ~300 curated images

• Specs: 2000 steps, 20 Epochs, Rank/Dim/Alpha 16, LR 0.0001

• Trainer: Civitai On-Site Cloud Trainer

I want to give a massive shoutout to everyone who supported my previous concept models! Because you guys supported my temporary Early Access releases (which are now completely free!), I was able to use that Buzz to offload this massive ~300-image dataset to the Civitai cloud trainer. My local hardware struggles with anything over 100 images, so your support directly funded this model and saved my GPU!

☕