Collected sources and patterns will appear here. Add from search or the patterns library.
FrameEmbedding -> List<QuantizedToken>
Predict high-order audio codebooks autoregressively within each frame using a small decoder conditioned on a base frame embedding.
Problem it solves
Modeling high-dimensional multi-codebook neural audio directly in a massive transformer scales poorly in sequence length.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.