Encoder-Free Vision-Language Models
Investigating architectures that unify visual and language understanding and generation directly in pixel space without separate encoders.
1 papers
Where this stands
The written synthesis of this thread is for subscribers. Subscribe.