Looking into it now, just out of curiosity and to stay up on things. Didn't realize it was such a breakthrough in power/open-weight/local-runnability. Which version you running? Seems there are 3 variants:
FP16 (full precision): ~54 GB → not feasible on most Macs
8-bit: ~27 GB → borderline
4-bit (quantized): ~13–18 GB → practical
Probably the full precision cuz you have a powerhouse workstation, I'm guessing. Seems with a lowly MacBook I'd be relegated to the lowest tier. Someday soonish, I'll be wanting to get a powerful box I can run good local stuff on. Not sure why, since I don't really do much personal stuff with them, but it just seems cool.
Thanks for sharing/entertaining my q's.
