mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-09-29 01:17:36 -05:00
* main : add option to save full output to session * split behavior into --session and --prompt-cache * restore original implementation with new names * PR comments * move the check for incompatible parameters to gpt_params_parse * Fix whitespace Co-authored-by: DannyDaemonic <DannyDaemonic@gmail.com> --------- Co-authored-by: DannyDaemonic <DannyDaemonic@gmail.com>