DiG: Scalable and Efficient Diffusion Models with Gated Linear Attention 【拡散モデル勉強会】DiG: Scalable and Efficient Diffusion Models with Gated Linear Attention by more read more