Highly recommend using my model, Janku, with my positive and negative embeddings, Lazy embeddings. I test all my LORAs with them!

Please share the content you make by tagging this LORA. Feel free to DM me too! I would love to see what you can create with this LORA!

Updated to v1.1 with increased quality - check how the knight action improved in the images!

Trigger word: aimindstyle

LORA strength: 1.0

Works well with a combination of: sketch, digital media artwork, traditional media or any other quality tags.

Check the knight picture for LORA strength examples. Don't go below 0.9 or the style doesn't show that well. All my images uploaded have my ComfyUI workflow in them, just import them in ComfyUI to replicate image 1:1.

I also do LORA commissions, but if the character is popular and don't have LORAs on here I will do it for free! See my profile or https://ko-fi.com/janxd for more details!

All credits go to deltax237 for his training guide and setting up the training data.

Trained on: https://www.deviantart.com/in-the-mind-of-ai 40 images

My training settings

Network Dimension Rank 32- Network Alpha Rank 16

Optimizer -> AdamW8, Cosine with Restarts, 3 cycles

Batch Size 4, Epochs 10, Repeats 5