Video LoRA Training for one-eye Calico Cat

WAN 2.2 video LoRA for my one-eyed calico cat with unique fur pattern, one of a-kind color mix goal: WAN 2.2 video LoRA file for WAN 2.2 base model to generate short, stylised clips of her in different scenes... (I have a ZIT image LoRA for my one-eyed calico cat, that is why I can use WAN2.2 I2V for much better consistency than T2V) dataset is 100-200 raw clips, each 3-10 seconds long (or even more) – Footage should show natural behaviours (walking, jumping, grooming, stretching, playing) from multiple angles in daylight >>>>>>>>>>>>>>> first step what would you suggest for video captioning ? also considering temporal relationship... frame extraction, resizing, tagging or captioning, feel free to recommend what is best in your option... for image LoRA I was using solotagger with Gemma 4 26B A4B what has much better results than JoyCaption BetaOne.. important: just keep the pipeline reproducible so I can 'update' the WAN2.2 video LoRA later if I will have new content... (please do also send your portfolio)

Project ID: 40518315