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 2024-01-22 11:25:11 -05:00
parent 5b982d0f8c
commit fcdf337d84
2 changed files with 3 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
llama_cpp/llama_cpp.py
View file

@ -188,6 +188,7 @@ LLAMA_TOKEN_TYPE_BYTE = 6
# LLAMA_FTYPE_MOSTLY_IQ2_XXS = 19, // except 1d tensors # LLAMA_FTYPE_MOSTLY_IQ2_XXS = 19, // except 1d tensors
# LLAMA_FTYPE_MOSTLY_IQ2_XS = 20, // except 1d tensors # LLAMA_FTYPE_MOSTLY_IQ2_XS = 20, // except 1d tensors
# LLAMA_FTYPE_MOSTLY_Q2_K_S = 21, // except 1d tensors # LLAMA_FTYPE_MOSTLY_Q2_K_S = 21, // except 1d tensors
# LLAMA_FTYPE_MOSTLY_Q3_K_XS = 22, // except 1d tensors
# LLAMA_FTYPE_GUESSED = 1024, // not specified in the model file # LLAMA_FTYPE_GUESSED = 1024, // not specified in the model file
# }; # };
@ -211,6 +212,7 @@ LLAMA_FTYPE_MOSTLY_Q6_K = 18
LLAMA_FTYPE_MOSTLY_IQ2_XXS = 19 LLAMA_FTYPE_MOSTLY_IQ2_XXS = 19
LLAMA_FTYPE_MOSTLY_IQ2_XS = 20 LLAMA_FTYPE_MOSTLY_IQ2_XS = 20
LLAMA_FTYPE_MOSTLY_Q2_K_S = 21 LLAMA_FTYPE_MOSTLY_Q2_K_S = 21
LLAMA_FTYPE_MOSTLY_Q3_K_XS = 22
LLAMA_FTYPE_GUESSED = 1024 LLAMA_FTYPE_GUESSED = 1024
# enum llama_rope_scaling_type { # enum llama_rope_scaling_type {

2
vendor/llama.cpp vendored

@ -1 +1 @@
Subproject commit 504dc37be8446fb09b1ede70300250ad41be32a2 Subproject commit 6f9939d119b2d004c264952eb510bd106455531e