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 2023-04-12 14:29:00 -04:00
parent c854c2564b
commit 005c78d26c
2 changed files with 2 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

1
llama_cpp/llama_cpp.py
View file

@ -114,6 +114,7 @@ LLAMA_FTYPE_ALL_F32 = ctypes.c_int(0)
LLAMA_FTYPE_MOSTLY_F16 = ctypes.c_int(1) # except 1d tensors LLAMA_FTYPE_MOSTLY_F16 = ctypes.c_int(1) # except 1d tensors
LLAMA_FTYPE_MOSTLY_Q4_0 = ctypes.c_int(2) # except 1d tensors LLAMA_FTYPE_MOSTLY_Q4_0 = ctypes.c_int(2) # except 1d tensors
LLAMA_FTYPE_MOSTLY_Q4_1 = ctypes.c_int(3) # except 1d tensors LLAMA_FTYPE_MOSTLY_Q4_1 = ctypes.c_int(3) # except 1d tensors
LLAMA_FTYPE_MOSTLY_Q4_1_SOME_F16 = ctypes.c_int(4) # tok_embeddings.weight and output.weight are F16
# Functions # Functions

2
vendor/llama.cpp vendored

@ -1 +1 @@
Subproject commit 8b679987cdce292ff36bd741f6715e4927e26f9b Subproject commit e7f6997f897a18b6372a6460e25c5f89e1469f1d