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

PR #4766

Browse source
This commit is contained in:
Jared Van Bortel 2024-01-10 11:29:04 -05:00
parent 3773e1afe7
commit 1eb8804c18
18 changed files with 2183 additions and 2081 deletions
Download patch file Download diff file Expand all files Collapse all files

17
llama.h
View file

@ -116,6 +116,12 @@ extern "C" {
LLAMA_ROPE_SCALING_MAX_VALUE = LLAMA_ROPE_SCALING_YARN,
};
enum llama_split_mode {
LLAMA_SPLIT_NONE = 0, // single GPU
LLAMA_SPLIT_LAYER = 1, // split layers and KV across GPUs
LLAMA_SPLIT_ROW = 2, // split rows across GPUs
};
typedef struct llama_token_data {
llama_token id; // token id
float logit; // log-odds of the token
@ -178,8 +184,15 @@ extern "C" {
struct llama_model_params {
int32_t n_gpu_layers; // number of layers to store in VRAM
int32_t main_gpu; // the GPU that is used for scratch and small tensors
const float * tensor_split; // how to split layers across multiple GPUs (size: LLAMA_MAX_DEVICES)
enum llama_split_mode split_mode; // how to split the model across multiple GPUs
// main_gpu interpretation depends on split_mode:
// LLAMA_SPLIT_NONE: the GPU that is used for the entire model
// LLAMA_SPLIT_ROW: the GPU that is used for small tensors and intermediate results
// LLAMA_SPLIT_LAYER: ignored
int32_t main_gpu;
// proportion of the model (layers or rows) to offload to each GPU, size: LLAMA_MAX_DEVICES
const float * tensor_split;
// Called with a progress value between 0.0 and 1.0. Pass NULL to disable.
// If the provided progress_callback returns true, model loading continues.