News for hackers. Stuff that matters.

A comment on High-Fidelity Simultaneous Speech-to-Speech Translation.

Re:High-Fidelity Simultaneous Speech-to-Speech Translation by nottorp on Friday July 04, 2025 @07:24AM (#44463535)
Is this deterministic or random like a LLM?
[ Parent | on HN ]
Re:High-Fidelity Simultaneous Speech-to-Speech Translation by l-m-z on Friday July 04, 2025 @10:45AM (#44465018)
Hibiki is an auto-regressive model with temperature based sampling so very similar to a LLM, generations are "random" and you can make them deterministic by fixing the RNG seed.
[ Parent | on HN ]