Announcement
Gradium's default text-to-speech model is claimed to deliver first audio in around 50ms
Gradium says the model scores highest for naturalness among sub-100ms models on Speko's public benchmarks.
TLDR
Gradium says its default text-to-speech model now produces its first audio in around 50 milliseconds. It also claims the highest naturalness score among sub-100ms models on Speko's public benchmarks.
Combined views
6.2K
2 Sources, first seen ago
