llama.cpp
fit : account for skipped NextN layers when setting n_gpu_layers
#31569
Merged
Merged
fit : account for skipped NextN layers when setting n_gpu_layers #31569
ggerganov
merged
3 commits into
master
from
aman/fit-nextn-offset
fit : account for skipped NextN layers when setting n_gpu_layers
159c23f2
CISC
approved these changes
on 2026-11-03
fit : pass hp_ngl_skip by reference like the other hparams
4e2a188d
JohannesGaessler
commented
on 2026-11-03
fit : add hp_ngl_skip to common_get_device_memory_data
08423882
JohannesGaessler
approved these changes
on 2026-11-05
Login to write a write a comment.
Login via GitHub