alt
Hacker News
trenchpilgrim
•
last Saturday at 9:24 PM
•
0 replies
•
view on HN
Certain chips can work with useful local models, but compatibility is far behind CUDA.
alt Hacker News