Processing 24T tokens for LLM training with 0 crashes (what made it possible) (daft.ai) 1 points by DISCURSIVE 1y ago ↗ HN
0 comments
[ 0.25 ms ] story [ 11.3 ms ] threadNo comments yet.