This paper demonstrates that image generation pretraining serves as a unified foundation for both visual creation and zero-shot understanding, rivaling domain-specific specialists across diverse 2D and 3D tasks. It proposes a paradigm shift where generative models act as generalist vision learners, establishing image generation as a universal interface for computer vision similar to text in LLMs.
Listen on: Spotify · Apple Podcasts · Amazon Music · YouTube
Follow the show: YouTube · Medium · X · LinkedIn · Bluesky · Threads · Instagram · TikTok · Facebook · Telegram