Retrocomicflux

🚀 Retro Comic Flux

Retro Comic Flux is a LoRA model that generates retro - style comic book images. It is trained on a curated dataset of public - domain images, offering unique and vivid comic - themed outputs.

🚀 Quick Start

Trigger Words

You should use c0m1c and comic book panel to trigger the image generation.

Recommended Strengths

The recommended strengths for using this model are in the range of 0.7 - 0.9.

Sampler Recommendation

USE DEIS SAMPLER FOR THE BEST RESULTS.

✨ Features

• Unique Training Data: Trained on a dataset of manually sourced, cropped, and enhanced public - domain images.
• Human - Figure Focus: Primarily good at generating human figures, especially when you describe the character and a backdrop scene.
• Color and Contrast Enhancement: Images in the training set were pre - processed in Photoshop to bring out vivid colors, reduce yellowing, and boost contrast, improving training quality.
• Flexible Prompting: You can explicitly state "empty speech bubble" in prompts, and the model responds well to natural language descriptions.

📦 Installation

The weights for this model are available in Safetensors format. You can Download them in the Files & versions tab.

📚 Documentation

Model Description

Retro Comic Flux is a Flux LoRA trained on a dataset consisting entirely of public domain images manually sourced, cropped and enhanced. I used ~50 images with 3 repeats, 15 epochs, and a 0.0001 learning rate. These were all captioned using Joy Caption Batch.

This is a very experimental model. I had a difficult time sourcing high - quality images of animals and vehicles in this style, so for now, it's primarily good with human figures. It works best when you describe the character and a scene for a backdrop. It struggles with hands and eyes sometimes. If something is really not in the dataset, it defaults to an ordinary AI illustration style.

I used very similar settings as I have with my previous LoRAs. Low learning rates with less repeats and good data work best for simple styles.

I was careful in my selection of images and pre - processed each image in Photoshop to bring out vivid colors, reduce the yellowing, and boost contrast. This helped me to maintain a consistent coloring and I believe it improved the training quality. When I did this, I also removed the speech from speech bubbles from some of the training data for more flexibility and to help Flux understand context. You can explicitly state "empty speech bubble" in your prompts, and it tends to work. If you want to dive in deeper, try using an image captioner to describe a style you want to emulate. The images have rich captions, and the model responds well to natural language.

Definitely expect a V2 of this LoRA soon. I have some ideas on how I can improve consistency, but I need to find more varied source images.

Trigger Words

• You should use c0m1c to trigger the image generation.
• You should use comic book panel to trigger the image generation.

📄 License

This model is licensed under the creativeml - openrail - m license.

Gallery