Autoregressive visual & audio models

训练方式

A learned codec is trained for reconstruction; the prior minimizes next-code cross-entropy, optionally conditioned on text, prior frames, semantic tokens, or coarse-scale codes.

推理运行方式

Codes are generated serially or in hierarchical/blockwise schedules, then decoded. Ordering determines latency and which dependencies are easy to model.

优势

• •Unified token-based modeling across modalities
• •Compatible with causal language-model infrastructure
• •Exact discrete likelihood for the chosen tokenization

权衡

• •Long code sequences create serial latency
• •Tokenizer artifacts bound output quality
• •A one-dimensional ordering can be awkward for spatial or multiscale structure

适用场景

• A strong discrete codec exists
• Cross-modal token modeling or continuation is important
• Hierarchical generation can meet latency goals

应避免或质疑的场景

• Long serial decoding misses the budget
• The codec loses critical perceptual detail
• A continuous diffusion representation is substantially simpler