Skip to main content
info

"Informed AI News" is an publications aggregation platform, ensuring you only gain the most valuable information, to eliminate information asymmetry and break through the limits of information cocoons. Find out more >>

Advanced AI Generates Multimodal Long Stories

SEED-Story crafts long, interleaved image-text tales. It utilizes a Multimodal Large Language Model (MLLM) to predict both text and visual tokens, transforming them into consistent images. A new attention sink boosts efficiency, allowing up to 25 sequences. StoryStream, a high-resolution dataset, aids in training and evaluation.

Full article>>