mirror of
https://github.com/likelovewant/ollama-for-amd.git
synced 2025-12-22 06:43:57 +00:00
If we try to use mmap when the model is larger than the system free space, loading is slower than the no-mmap approach.
29 KiB
29 KiB