We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
f6fbdf3
There was an error while loading. Please reload this page.
Better user experience with carefully tuned inference-time hyper parameters in Omni Full-Duplex mode. (length penalty 1.05, 1.1, 1.2)
5169e63
Improve torch.compile behavior among different modes.
40e7da2
Better user experience with carefully tuned inference-time hyper parameters in Omni Full-Duplex mode.
501e65e
Add torch.compile to accelerate inference
