README.md

DAC — AEmotionStudio mirror

Mirrors of Descript's Descript Audio Codec checkpoints, used by Tessera v2 Latent mode (MAESTRO).

Upstream: https://github.com/descriptinc/descript-audio-codec
License: MIT.

Files

• weights_44khz_8kbps.pth — audiotools BaseModel.save wrapper ({state_dict, metadata} — needed by DAC.load).
• weights_24khz_8kbps.pth — audiotools BaseModel.save wrapper ({state_dict, metadata} — needed by DAC.load).
• weights_16khz_8kbps.pth — audiotools BaseModel.save wrapper ({state_dict, metadata} — needed by DAC.load).

Loading

import dac
model = dac.model.DAC.load('weights_44khz_8kbps.pth')
model.eval()