| ▲ | Fusing a 27B ternary LLM's whole decode step into one CUDA kernel(twitter.com) | |
| 3 points by Jr23_xd 5 hours ago | 1 comments | ||
| ▲ | genxy an hour ago | parent [-] | |
[dead] | ||
| ▲ | Fusing a 27B ternary LLM's whole decode step into one CUDA kernel(twitter.com) | |
| 3 points by Jr23_xd 5 hours ago | 1 comments | ||
| ▲ | genxy an hour ago | parent [-] | |
[dead] | ||