According to KREA AI (@krea_ai), the company has introduced trainers for the Qwen-2512 and Z-Image models, enabling users to train LoRA (Low-Rank Adaptation) modules tailored to specific needs. These LoRAs can now be seamlessly integrated and utilized within the Krea Image platform, allowing for customized AI image generation workflows. This development provides businesses and creators with expanded flexibility in fine-tuning image models for unique datasets and creative requirements, opening up new opportunities for personalized visual content and vertical applications in AI-powered design. Source: KREA AI Twitter (https://twitter.com/krea_ai/status/2009286474730996026)