Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
NightlyDev
on March 5, 2025
|
parent
|
context
|
favorite
| on:
Apple M3 Ultra
The full deepseek R1 model needs more memory than 512GB. The model is 720GB alone. You can run a quantized version on it, but not the full model.
summarity
on March 5, 2025
[–]
You can chain multiple Mac Studios using exo for inference, you'd "only" need two of these. There's a bottleneck in the RMA speed over TB5, but this may not matter as much for a MoE model.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
