| ▲ | Show HN: TurboGPT: train 22KiB transformer in 13s(github.com) | |
| 12 points by lostmsu 2 hours ago | 1 comments | ||
Inspired by minGPT for home experiments. Requires CUDA 13.4; build script is Windows only. | ||
| ▲ | jey an hour ago | parent [-] | |
At that scale it seems like you could just solve the KKT conditions directly. Exaggerating but only a little | ||