ConceptAI & ML
Modality-specific positional encodings
Learned special embeddings for each modality used to tag features from audio or video to allow the network to distinguish them.
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.
Multimodal data1Positional encoding1
Lecturers
1Connections
2The lectures that define it or assume it, and the concepts next to it.