alihan/llama-cpp-python
1
0
Fork 0
mirror of https://github.com/abetlen/llama-cpp-python.git synced 2023-09-07 17:34:22 +03:00
Code Issues Packages Projects Releases Wiki Activity

Update llama.cpp

Browse Source
This commit is contained in:
Andrei Betlen
2023-04-26 20:00:54 -04:00
parent cbd26fdcc1
commit 9339929f56
2 changed files with 9 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
llama_cpp/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_2 = ctypes.c_int(5) # except 1d tensors
LLAMA_FTYPE_MOSTYL_Q4_3 = ctypes.c_int(6) # except 1d tensors LLAMA_FTYPE_MOSTYL_Q4_3 = ctypes.c_int(6) # except 1d tensors
LLAMA_FTYPE_MOSTYL_Q8_0 = ctypes.c_int(7) # except 1d tensors LLAMA_FTYPE_MOSTYL_Q8_0 = ctypes.c_int(7) # except 1d tensors
LLAMA_FTYPE_MOSTYL_Q5_0 = ctypes.c_int(8) # except 1d tensors
LLAMA_FTYPE_MOSTYL_Q5_1 = ctypes.c_int(9) # except 1d tensors
# Functions # 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.argtypes = [llama_context_p]
_lib.llama_get_kv_cache_token_count.restype = c_int _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) # 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: def llama_get_state_size(ctx: llama_context_p) -> c_size_t:

2
vendor/llama.cpp vendored

Submodule vendor/llama.cpp updated: 4afcc37869...0b2da20538