Code and results for: Lossless Entropy Coding of Neural Speech Tokens: Language-Model Tokenisers, Dedicated Entropy Models and Resilience to Packet Loss

Code, result files, training curves and file lists for the article "Lossless Entropy Coding of Neural Speech Tokens", submitted to IEEE/ACM Transactions on Audio, Speech, and Language Processing. Includes scripts for token extraction from six neural codecs, the TD-Prior entropy model, range coding, static references, evaluation of the EnCodec entropy model, segmented coding and quality evaluation.