vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-10 17:14:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

still refactoring

Browse source
This commit is contained in:
Concedo 2023-04-01 11:56:34 +08:00
parent 6e6125ebdb
commit 085a9f90a7
11 changed files with 72 additions and 261 deletions
Download patch file Download diff file Expand all files Collapse all files

10
llama_adapter.cpp
View file

@ -19,7 +19,7 @@ static gpt_params params;
static int n_past = 0;
static int n_threads = 4;
static int n_batch = 8;
static std::string model;
static std::string modelname;
static llama_context *ctx;
static std::vector<llama_token> last_n_tokens;
static std::vector<llama_token> current_context_tokens;
@ -32,7 +32,7 @@ bool llama_load_model(const load_model_inputs inputs, FileFormat in_file_format)
n_threads = inputs.threads;
n_batch = inputs.batch_size;
model = inputs.model_filename;
modelname = inputs.model_filename;
ctx_params.n_ctx = inputs.max_context_length;
ctx_params.n_parts = inputs.n_parts_overwrite;
@ -44,16 +44,16 @@ bool llama_load_model(const load_model_inputs inputs, FileFormat in_file_format)
if (file_format == FileFormat::GGML || file_format == FileFormat::GGHF)
{
ctx = legacy_llama_init_from_file(model.c_str(), ctx_params);
ctx = legacy_llama_init_from_file(modelname.c_str(), ctx_params);
}
else
{
ctx = llama_init_from_file(model.c_str(), ctx_params);
ctx = llama_init_from_file(modelname.c_str(), ctx_params);
}
if (ctx == NULL)
{
fprintf(stderr, "%s: error: failed to load model '%s'\n", __func__, model.c_str());
fprintf(stderr, "%s: error: failed to load model '%s'\n", __func__, modelname.c_str());
return false;
}