▲
Hacker News
new
|
best
|
ask
|
show
|
jobs
points by
lostmsu
2 days ago | 0 comments
> Small: 594 billion parameters
cries in GPU racks
[–]
colingauvin
2 days ago
You could probably run a Q2-Q3 with 2x DGX Sparks. So...$12k at today's prices. Plenty of folks are running DS 4.1 Flash with Strata and some GPU + bulk in system memory. Maybe more like $3k-$4k? But speeds come down quite a bit.
[–]
joshspankit
2 days ago
Q2. Woof.
[–]
colingauvin
1 day ago
It actually benchmarks quite well. Honestly not practically worse than the Q4 native weights.
[–] colingauvin 2 days ago
[–] joshspankit 2 days ago
[–] colingauvin 1 day ago