Congratulations to @DecagonAI on the launch of Chord.
Trained and served on Modal, Chord sees low time-to-first-audio even during massive swings in concurrency.
Decagon launches Chord, a speech model trained and served on Modal
AISummary
Decagon Labs has launched Chord, a new speech model, and Modal says it was trained and served on its platform. Modal reports low time-to-first-audio even during large swings in concurrency. Decagon says post-training preserved pace and voice naturalness and raised resolution rates in every deployment.
Post on XView on X
@modal
Meet Chord, our new speech model by the Decagon Labs team. Post-trained to preserve pace and voice naturalness, Chord has increased resolution rates in every deployment it was used in.
Source: Modal · x.comPublished · added here