Adobe inc. (20250117972). MODALITY SPECIFIC LEARNABLE ATTENTION FOR MULTI-CONDITIONED DIFFUSION MODELS
MODALITY SPECIFIC LEARNABLE ATTENTION FOR MULTI-CONDITIONED DIFFUSION MODELS
Organization Name
Inventor(s)
Hareesh Ravi of San Jose CA US
Aashish Kumar Misraa of Santa Clara CA US
Ajinkya Gorakhnath Kale of San Jose CA US
MODALITY SPECIFIC LEARNABLE ATTENTION FOR MULTI-CONDITIONED DIFFUSION MODELS
This abstract first appeared for US patent application 20250117972 titled 'MODALITY SPECIFIC LEARNABLE ATTENTION FOR MULTI-CONDITIONED DIFFUSION MODELS
Original Abstract Submitted
a method, apparatus, non-transitory computer readable medium, and system for image generation include encoding a text prompt to obtain a text embedding. an image prompt is encoded to obtain an image embedding. cross-attention is performed on the text embedding and then on the image embedding to obtain a text attention output and an image attention output, respectively. a synthesized image is generated based on the text attention output and the image attention output.