← Back to context Comment by pinstripes 4 months ago Trying to scale text inference to 1 million tok/s on cheap hardware. 0 comments pinstripes Reply No comments yet Contribute on Hacker News ↗
No comments yet
Contribute on Hacker News ↗