関連ワード:  text to image diffusion models are zero shot classifiers github  
結果 : text to image diffusion models are zero shot classifiers github