3 ms·
My hope is that we can get to the point I can run good-enough models on local hardware before they are all ad laden.
by ModernMech 7mo ago
My hope is that we can get to the point I can run good-enough models on local hardware before they are all ad laden.
- aurareturn 7mo agoYou can already do this. Get an M5 Max with 129GB model and run a Qwen model. Or with for an M5 Ultra with 512GB of RAM and run Deepseek/Kimi/larger Qwen.