search

descriptinc / descript-audio-codec

State-of-the-art audio codec with 90x compression factor. Supports 44.1kHz, 24kHz, and 16kHz mono/stereo audio.
https://descript.notion.site/Descript-Audio-Codec-11389fce0ce2419891d6591a68f814d5
MIT License
1.2k stars 111 forks source link

Implement "Residual Quantization with Implicit Neural Codebooks" #90

Open Jourdelune opened 3 months ago

Jourdelune commented 3 months ago

Hello, the paper "Residual Quantization with Implicit Neural Codebooks" (https://arxiv.org/abs/2401.14732) improve RVQ. I am sharing it here if you want to improve DAC.

Jourdelune commented 2 months ago

There is an implementation here: https://github.com/lucidrains/vector-quantize-pytorch/commit/f47b7db0ed7f090fc736731103a6b84d9e3b6265