Up to 8.3 times cheaper inference, 6 times more context, and 10 times the speed.
1 comment
quantbagelabout 3 hours ago
we distilled GPT-6 Astra, into Nvidia's DreamZero world model. Astra is state-of-the-art on most robotics benchmarks, but its too slow. On the Dream Zero backbone its much faster! Of course, there's some degradation, but the speed improvement means it's actually usable on high-end robots to amateur arms.
Read the full thread on Hacker News →
Related stories
- Hacker News · 186 points · 1 day ago
- Hacker News · 1 points · 4 days ago
- Hacker News · 3 points · 11 days ago
- Hacker News · 99 points · 1 day ago
- DEV Community · 8 points · 21 days ago
- DEV Community · 3 points · 9 days ago