15 ms·
Cerebras does not share the quantization of the models so you don't know if you're getting real K3 or k3 lite or something else.
by sergiotapia 3mo ago
Cerebras does not share the quantization of the models so you don't know if you're getting real K3 or k3 lite or something else.
- codexon 3mo agoIt most likely will be quantized. A cerebras wafer only has 44gb ram, and linking them together vastly reduces the speedup.