Processing 24T tokens for LLM training with 0 crashes (what made it possible) (daft.ai) 1 points by DISCURSIVE 1y ago ↗ HN
0 comments
[ 3.2 ms ] story [ 11.9 ms ] threadNo comments yet.