BWhat changed
Modeled interleaved text/audio history with a zeroth-codebook backbone and smaller residual-codebook decoder, subsampling decoder training frames and evaluating contextual pronunciation/prosody.
Scientific scope
Modeled interleaved text/audio history with a zeroth-codebook backbone and smaller residual-codebook decoder, subsampling decoder training frames and evaluating contextual pronunciation/prosody. Inherits RQTransformer, Llama-style architecture and Mimi codec from Moshi. Acquired delta is zeroth-codebook backbone/small decoder split, decoder-frame subsampling and contextual evaluation integration, not first hierarchical audio transformer. Technical post admits deficient conversational structure and a context-sensitive human-prosody gap.
Evidence
The linked original sources describe this contribution and its evidence limits. WRS has not independently reconstructed the full scientific claims. Inherits RQTransformer, Llama-style architecture and Mimi codec from Moshi. Acquired delta is zeroth-codebook backbone/small decoder split, decoder-frame subsampling and contextual evaluation integration, not first hierarchical audio transformer. Technical post admits deficient conversational structure and a context-sensitive human-prosody gap.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction