llama.cpp
llama: (proposal) propagating the results of `graph_compute` to the user interface
#9525
Merged
Merged
llama: (proposal) propagating the results of `graph_compute` to the user interface #9525
ggerganov
merged
7 commits into
ggml-org:master
from
Xarbirus:ggml_status_to_user
vignesh1507
approved these changes
on 2024-10-06
llama: propagating the results of `graph_compute` to the user interface
5e354e3c
llama: reverting kv_cache in case of failed compute
47018932
llama: `llama_kv_cache_state` was removed, only the result of `llama_…
acb95283
llama: restore a kv_cache in case of failed computation
0026c810
llama: correct reverting of the entire batch.
ee599f90
llama: updated comments
0638c448
llama : add comments about KV cache state after error
9ef5d089
Login to write a write a comment.
Login via GitHub