undefined | Better HN

0 pointstatef3mo ago0 comments

Hypura reads tensor weights from the GGUF file on NVMe into RAM/GPU memory pools, then compute happens entirely in RAM/GPU.

There is no writing to SSDs on inference with this architecture.

0 comments

1 comments · 1 top-level

embedding-shape3mo ago

Even if there was a ton of writing, I'm not sure where NVMe even comes in the picture, write durability is about the flash cells on SSDs, nothing to do with the interface, someone correct me if I'm wrong.

j / k navigate · click thread line to collapse