llama.cpp
llama.cpp : split llama_context_params into model and context params
#3301
Merged
Merged
llama.cpp : split llama_context_params into model and context params #3301
ggerganov
merged
17 commits into
master
from
llama-model-params
llama.cpp : split llama_context_params into model and context params
cf1f8059
fix metal build
39f4afac
fix freq_base/scale default to model value
f28e4953
llama-bench : keep the same model between tests when possible
96f6dcde
move n_threads to llama_context_params, add n_threads_batch
7f953792
fix mpi build
d41b53ca
remove kv_size(), cuda scratch fixes
ceb18e44
remove low-vram option
92fb8ab5
ggerganov
approved these changes
on 2023-09-22
add n_threads_batch to system info, refactor to get_system_info()
a6084cc7
add documentation about --threads-batch to the READMEs
34696841
llama-bench fix
e5afe420
main : fix rope freq/scale warning
0e9ed7f8
llama.cpp : add llama_get_model
8f5b0eaa
Merge remote-tracking branch 'origin/master' into llama-model-params
65b83f37
remove duplicated ctx/model functions
5659391b
slaren
commented
on 2023-09-28
cuda : print total VRAM used
17e841ac
Merge remote-tracking branch 'origin/master' into llama-model-params
c8a9658e
Reviewers
Assignees
No one assigned
Labels
high priority
breaking change
refactoring
Milestone
No milestone
Login to write a write a comment.
Login via GitHub