mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-05-07 09:34:07 +00:00
* speculative : update default params * speculative : do not discard the last drafted token
176 KiB
176 KiB