llama.cpp
server: Fixed speculative decoding stats to use `#accepted \ #tested` rather than `#accepted \ #drafted`
#14104
Merged
Merged
server: Fixed speculative decoding stats to use `#accepted \ #tested` rather than `#accepted \ #drafted` #14104
jukofyork
merged
1 commit into
ggml-org:master
from
spec-timings-fix
Fixed spec timings to: tested/accepted instead of drafted/accepted
0bda7877
ggerganov
approved these changes
on 2025-06-10
Login to write a write a comment.
Login via GitHub