higgs_audio/constants.py

AUDIO_IN_TOKEN = "<|AUDIO|>"
AUDIO_OUT_TOKEN = "<|AUDIO_OUT|>"
EOS_TOKEN = "<|end_of_text|>"