タイプ
LORA
投稿日時
2025-08-29
基本モデル
Wan Video 14B t2v










lora of a fictional woman named madison
put owl in the negative prompt if it starts generating owls
purpose of this is to catalog progress as it improves in quality
V001: Issues with lighting on jawline. need to prune dataset and retrain. also grid texture appears in background. probably need to gen data images at higher res and scale down.
V002: enabled flip augmentation and seems to have fixed the majority of the gridlines issue
V003: ran images through an upscaler and completely got rid of the grid lines
wan2.2 14b v001: no keyword needed.
1.転載モデルは学習・共有目的のみで使用し、著作権は原作者に帰属します
2.モデルの認証は公式チャンネルでご連絡ください。クリエイターの権利保護に努めています クリックして認証
