| # Notice | |
| This repository is built from BayLing-Duplex and uses upstream GLM-4-Voice components. | |
| The model uses upstream components including: | |
| - the upstream checkpoint format and tokenizer files; | |
| - the upstream speech tokenizer; | |
| - the upstream speech decoder based on CosyVoice components. | |
| Code under `bayling_duplex_runtime/cosyvoice` contains files derived from CosyVoice and related components with Apache-2.0 headers. Keep the original source headers when redistributing or modifying those files. | |
| For upstream GLM-4-Voice information, see: | |
| - https://github.com/THUDM/GLM-4-Voice | |
| - https://github.com/THUDM/GLM-4 | |