vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-05-22 03:10:03 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

Merge commit '9230dbe2c7' into concedo_experimental

Browse source 
# Conflicts:
#	ggml/src/ggml-cpu/CMakeLists.txt
#	src/llama-graph.cpp
#	tools/server/README.md
This commit is contained in:
Concedo 2025-06-21 00:01:29 +08:00
commit c16d672ce4
29 changed files with 2666 additions and 2457 deletions
Download patch file Download diff file Expand all files Collapse all files

26
common/arg.cpp
View file

@ -3212,6 +3212,32 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
params.speculative.model.path = value;
}
).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_MODEL_DRAFT"));
add_opt(common_arg(
{"-ctkd", "--cache-type-k-draft"}, "TYPE",
string_format(
"KV cache data type for K for the draft model\n"
"allowed values: %s\n"
"(default: %s)",
get_all_kv_cache_types().c_str(),
ggml_type_name(params.speculative.cache_type_k)
),
[](common_params & params, const std::string & value) {
params.speculative.cache_type_k = kv_cache_type_from_str(value);
}
).set_env("LLAMA_ARG_CACHE_TYPE_K_DRAFT"));
add_opt(common_arg(
{"-ctvd", "--cache-type-v-draft"}, "TYPE",
string_format(
"KV cache data type for V for the draft model\n"
"allowed values: %s\n"
"(default: %s)",
get_all_kv_cache_types().c_str(),
ggml_type_name(params.speculative.cache_type_v)
),
[](common_params & params, const std::string & value) {
params.speculative.cache_type_v = kv_cache_type_from_str(value);
}
).set_env("LLAMA_ARG_CACHE_TYPE_V_DRAFT"));
add_opt(common_arg(
{"-mv", "--model-vocoder"}, "FNAME",