Author
Hong Huang
Papers that credit Hong Huang in their source bylines. Matching names may refer to different researchers.
1 results
One model learns faster image generation and understanding
A two-stage method cuts decoding steps while improving image and captioning results for a discrete multimodal model.