[–] thomasthelliez 2mo ago ↗ A local model generating 20 tokens/sec today could potentially reach 40 tokens/sec in many scenarios.
1 comment
[ 4.3 ms ] story [ 9.5 ms ] thread