| Property | Value |
|---|---|
| License | CreativeML OpenRAIL-M |
| Base Model | black-forest-labs/FLUX.1-dev |
| Training Images | 15 Hi-RES Images |
| Optimal Dimensions | 1024x1024, 768x1024 |
What is Ton618-Only-Stickers-Flux-LoRA?
Ton618-Only-Stickers-Flux-LoRA is a specialized LoRA model designed for generating sticker-style illustrations with consistent white backgrounds. Built on the FLUX.1-dev base model, it utilizes advanced training parameters including AdamW optimizer and constant LR scheduling to produce high-quality sticker designs.
Implementation Details
The model employs sophisticated training parameters with a network dimension of 64 and network alpha of 32. It was trained across 15 epochs using carefully curated high-resolution images, with noise offset of 0.03 and multires noise discount of 0.1.
- Utilizes florence2-en labeling for natural language processing
- Implements 24 repeat steps with 2.9K total steps
- Features multires noise iterations of 10
Core Capabilities
- Generation of cohesive sticker designs with white backgrounds
- Optimal performance at 1024x1024 resolution
- Support for various design elements including text integration
- Consistent style maintenance across different subjects
Frequently Asked Questions
Q: What makes this model unique?
This model specializes in creating sticker-style illustrations with consistent white backgrounds and defined borders, trained specifically on high-resolution images with careful parameter tuning.
Q: What are the recommended use cases?
The model is ideal for creating digital stickers, merchandise designs, and illustration assets that require clean, white backgrounds and consistent styling. It works best with the trigger word "Only Sticker" in prompts.