vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-12 09:59:41 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 3

Merge branch 'upstream' into concedo_experimental

Browse source 
# Conflicts:
#	.github/workflows/bench.yml
#	.github/workflows/build.yml
#	.github/workflows/python-check-requirements.yml
#	README.md
#	docs/backend/SYCL.md
#	flake.lock
#	ggml/CMakeLists.txt
#	ggml/src/kompute-shaders/op_rope_f16.comp
#	ggml/src/kompute-shaders/op_rope_f32.comp
#	ggml/src/kompute-shaders/rope_common.comp
This commit is contained in:
Concedo 2024-08-14 22:25:43 +08:00
commit e8de0af3ec
18 changed files with 1326 additions and 101 deletions
Download patch file Download diff file Expand all files Collapse all files

7
include/llama.h
View file

@ -95,13 +95,10 @@ extern "C" {
LLAMA_VOCAB_PRE_TYPE_CODESHELL = 22,
};
// note: these values should be synchronized with ggml_rope
// TODO: maybe move this enum to ggml.h (ggml_rope_type)
enum llama_rope_type {
LLAMA_ROPE_TYPE_NONE = -1,
LLAMA_ROPE_TYPE_NORM = 0,
LLAMA_ROPE_TYPE_NEOX = 2,
LLAMA_ROPE_TYPE_GLM = 4,
LLAMA_ROPE_TYPE_NORM = 0,
LLAMA_ROPE_TYPE_NEOX = GGML_ROPE_TYPE_NEOX,
};
enum llama_token_type { //TODO: remove, required until per token attributes are available from GGUF file