DeepX
When Local AI Couldn’t Run, I Reconsidered Low-Power Inference Chips
DeepX is developing a chip that could run 100-billion-parameter models below 5W, promising for edge devices but still too early to buy for typical ind
1 article tagged with this topic
DeepX is developing a chip that could run 100-billion-parameter models below 5W, promising for edge devices but still too early to buy for typical ind