llama.cpp
[Tensor Parallel] Fix recurrent state serialization for partial reads and writes
#22362
Merged
Merged
[Tensor Parallel] Fix recurrent state serialization for partial reads and writes #22362
JohannesGaessler
merged
1 commit into
ggml-org:master
from
gaugarg-nv:fix_cache_serialization
Fix recurrent state serialization for partial reads and writes
92f00411
gaugarg-nv
changed the title 
JohannesGaessler
approved these changes
on 2026-04-26
CISC
approved these changes
on 2026-04-26
Login to write a write a comment.
Login via GitHub