github/gpt4all
1
0
Fork 0
mirror of https://github.com/nomic-ai/gpt4all.git synced 2025-09-06 11:00:48 +00:00
Code Issues Packages Projects Releases Wiki Activity

expose n_gpu_layers parameter of llama.cpp (#1890)

Browse Source 
Also dynamically limit the GPU layers and context length fields to the maximum supported by the model.

Signed-off-by: Jared Van Bortel <jared@nomic.ai>
This commit is contained in:
Jared Van Bortel
2024-01-31 14:17:44 -05:00
committed by GitHub
parent f549d5a70a
commit 061d1969f8
31 changed files with 381 additions and 157 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
gpt4all-backend/gptj.cpp
View File

@@ -672,8 +672,9 @@ GPTJ::GPTJ()
d_ptr->modelLoaded = false;
}
size_t GPTJ::requiredMem(const std::string &modelPath, int n_ctx) {
size_t GPTJ::requiredMem(const std::string &modelPath, int n_ctx, int ngl) {
(void)n_ctx;
(void)ngl;
gptj_model dummy_model;
gpt_vocab dummy_vocab;
size_t mem_req;
@@ -681,8 +682,9 @@ size_t GPTJ::requiredMem(const std::string &modelPath, int n_ctx) {
return mem_req;
}
bool GPTJ::loadModel(const std::string &modelPath, int n_ctx) {
bool GPTJ::loadModel(const std::string &modelPath, int n_ctx, int ngl) {
(void)n_ctx;
(void)ngl;
std::mt19937 rng(time(NULL));
d_ptr->rng = rng;