Design teams can build controlled image workflows and fine-tune the base model while evaluating its noncommercial and commercial licensing requirements. Core capabilities include Text-to-image generation, Instruction-based image editing, Masked edits, and Multi-reference conditioning.
Public code or model weights are linked from the project, with setup instructions for running or adapting the system. Additional characteristics include Structured VGL prompting, Typography and texture generation, Photographic style modes, and Non-distilled weights for fine-tuning.
