baalajimaestro/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

Update llama.cpp

Browse source
This commit is contained in:
Andrei Betlen 2024-02-05 13:26:50 -05:00
parent 7467f129e5
commit 3553b14670
2 changed files with 3 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
llama_cpp/llama_cpp.py
View file

@ -445,7 +445,7 @@ class llama_model_params(Structure):
# uint32_t n_batch; // prompt processing maximum batch size # uint32_t n_batch; // prompt processing maximum batch size
# uint32_t n_threads; // number of threads to use for generation # uint32_t n_threads; // number of threads to use for generation
# uint32_t n_threads_batch; // number of threads to use for batch processing # uint32_t n_threads_batch; // number of threads to use for batch processing
# int8_t rope_scaling_type; // RoPE scaling type, from `enum llama_rope_scaling_type` # int32_t rope_scaling_type; // RoPE scaling type, from `enum llama_rope_scaling_type`
# // ref: https://github.com/ggerganov/llama.cpp/pull/2054 # // ref: https://github.com/ggerganov/llama.cpp/pull/2054
# float rope_freq_base; // RoPE base frequency, 0 = from model # float rope_freq_base; // RoPE base frequency, 0 = from model
@ -502,7 +502,7 @@ class llama_context_params(Structure):
("n_batch", c_uint32), ("n_batch", c_uint32),
("n_threads", c_uint32), ("n_threads", c_uint32),
("n_threads_batch", c_uint32), ("n_threads_batch", c_uint32),
("rope_scaling_type", c_int8), ("rope_scaling_type", c_int32),
("rope_freq_base", c_float), ("rope_freq_base", c_float),
("rope_freq_scale", c_float), ("rope_freq_scale", c_float),
("yarn_ext_factor", c_float), ("yarn_ext_factor", c_float),

2
vendor/llama.cpp vendored

@ -1 +1 @@
Subproject commit 191221178f51b6e81122c5bda0fd79620e547d07 Subproject commit 78b00dda6c0d62c34f5371d47718defff6ed2b22