ConceptAI & ML
CogVLM
A vision-language model that adds vision expert parameters to the backbone while keeping original language model weights intact, achieving strong multi-modal performance without reducing language capability.
Know first
Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.
Backbone networks1Feed-forward layers1QKV matrix1
Lecturers
1Connections
1The lectures that define it or assume it, and the concepts next to it.