Hacker News new | past | comments | ask | show | jobs | submit
heroiccocoa | on: DeepSeek-R1
It's just a question of having enough VRAM+RAM to fit the model into memory.