Two Channel (Stream) Speech Prompt using API

[jlian2]

Due diligence

• I have done my due diligence in trying to find the answer myself.

Topic

The PyTorch implementation

Question

I am doing some inference experiments. I am curious if I can feed two channel speech into model and get response (something the same as dGSLM) without human-machine interaction.

[jlian2]

I meant it feels to me that moshi is only able to accept user input and return agent output. But it is not able to accept both user and agent input, although it is mentioned in section 3.4.3.