LLM inference FHE homomorphic encryption: 1s/token interactive, 6min/token full

Posted 4 hours ago by vkaufmann
1 points

https://forums.developer.nvidia.com/t/llm-inference-under-ckks-fhe-on-one-dgx-spark-1-s-token-interactive-and-a-full-every-layer-encrypted-run/381842

1 comments

Loading..