vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

llama/ggml: add LLM training support

Browse source 
more compact progress bar

refactor: llama_prepare_sbatch/ubatch

llama_save_model_to_file

gqa_mode arg for repeat_back

llama_opt_param_filter

ggml_graph_dup force_grads

refactor ggml_opt, fix test-opt
This commit is contained in:
Johannes Gäßler 2024-11-17 14:58:51 +01:00
parent a5203b4465
commit c25557362a
26 changed files with 1294 additions and 339 deletions
Download patch file Download diff file Expand all files Collapse all files

4
src/llama-quant.cpp
View file

@ -512,7 +512,7 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std::
nthread = std::thread::hardware_concurrency();
}
// mmap consistently increases speed Linux, and also increases speed on Windows with
// mmap consistently increases speed on Linux, and also increases speed on Windows with
// hot cache. It may cause a slowdown on macOS, possibly related to free memory.
#if defined(__linux__) || defined(_WIN32)
constexpr bool use_mmap = true;
@ -522,7 +522,7 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std::
llama_model_kv_override * kv_overrides = nullptr;
if (params->kv_overrides) {
auto v = (std::vector<llama_model_kv_override>*)params->kv_overrides;
auto * v = (std::vector<llama_model_kv_override>*)params->kv_overrides;
kv_overrides = v->data();
}