A new LoRA (Low-Rank Adaptation) model, named Eyes Direction LoRA, has been released for StableDiffusion, offering precise control over the direction of eyes in generated images. This tool, developed by eric-venti-seeds, utilizes a red dot to indicate the desired gaze direction, ensuring more accurate and consistent results compared to previous methods. The model is based on the FLUX.2 and Klein 9b architectures and is available on Hugging Face. AI
IMPACT Enhances user control over image generation, enabling more specific artistic direction for character-focused visuals.
RANK_REASON This is a release of a fine-tuning model (LoRA) for an existing image generation platform, not a core model release from a frontier lab.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →