Skip to content

vllm_omni.model_executor.models.nemotron_voicechat.duplex

Native Full-Duplex integration for NVIDIA NemotronLabs VoiceChat.

Modules:

Name Description
data_plane

Output projection for Nemotron VoiceChat's independent channels.

input

80 ms PCM packetization for Nemotron VoiceChat native duplex.

runtime

Engine policy for Nemotron VoiceChat's frame-locked duplex timeline.

serving_adapter

Realtime serving adapter for Nemotron VoiceChat native duplex.