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

Update llama.cpp

Browse source
This commit is contained in:
Andrei Betlen 2023-04-26 20:00:54 -04:00 • committed by Don Mahurin
parent 656190750d
commit 66ad132575
1 changed files with 8 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

8
examples/llama_cpp.py
View file

@ -120,6 +120,8 @@ LLAMA_FTYPE_MOSTLY_Q4_1_SOME_F16 = ctypes.c_int(
LLAMA_FTYPE_MOSTLY_Q4_2 = ctypes.c_int(5) # except 1d tensors
LLAMA_FTYPE_MOSTLY_Q4_3 = ctypes.c_int(6) # except 1d tensors
LLAMA_FTYPE_MOSTLY_Q8_0 = ctypes.c_int(7) # except 1d tensors
LLAMA_FTYPE_MOSTLY_Q5_0 = ctypes.c_int(8) # except 1d tensors
LLAMA_FTYPE_MOSTLY_Q5_1 = ctypes.c_int(9) # except 1d tensors
# Functions
@ -210,6 +212,12 @@ def llama_get_kv_cache_token_count(ctx: llama_context_p) -> c_int:
_lib.llama_get_kv_cache_token_count.argtypes = [llama_context_p]
_lib.llama_get_kv_cache_token_count.restype = c_int
# Sets the current rng seed.
def llama_set_rng_seed(ctx: llama_context_p, seed: c_int):
return _lib.llama_set_rng_seed(ctx, seed)
_lib.llama_set_rng_seed.argtypes = [llama_context_p, c_int]
_lib.llama_set_rng_seed.restype = None
# Returns the size in bytes of the state (rng, logits, embedding and kv_cache)
def llama_get_state_size(ctx: llama_context_p) -> c_size_t: