feat: implement NVIDIA Cosmos-Transfer2.5-2b video-to-video style transfer support with base64 video encoding and NVCF status polling

fix: update Hugging Face Space targets to active public repositories and correct predict endpoint arguments to prevent api parameter schema mismatch

fix: pin huggingface_hub to <0.26.0 to fix Gradio import crash for HfFolder

style: remove local Engine & Parameters panel and hardware status warnings; convert VideoGen Studio to be 100% cloud-powered via InvictaTill AI and Hugging Face Cloud Spaces

feat: complete local integration of AI mind codebase (brain, memory, etc.) in text2video, supporting both Direct Codebase execution and Remote HTTP fallback

fix: replace AutoPipelineForText2Video with DiffusionPipeline and add custom AI Mind Settings to prompt composer

fix: change space metadata colors to supported values (purple/blue) for HF validation

feat: upgrade interface with InvictaTill AI Prompt Enhancer, style presets, custom premium UI, and memory offload optimization