AI Toolkit adds Krea 2 LoRA support

Providing day-zero training support for creative models like Krea 2 reinforces the open-source AI Toolkit as the industry-standard workbench for fine-tuning diffusion models. This rapid integration ensures developers aren't locked into proprietary platform interfaces for model customization.

• –Day-zero support allows developers to bypass Krea's native web-based training tool in favor of local or customized cloud training pipelines
• –Krea 2's focus on aesthetic coherence and style references makes it highly desirable for custom fine-tunes, especially compared to more literal prompt-following models
• –Local training via AI Toolkit offers significant cost and privacy advantages over cloud-based subscription APIs for high-volume creator workflows
• –This update demonstrates the agility of open-source fine-tuning tooling to keep pace with rapid foundation model releases