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

llama : accept a list of devices to use to offload a model (#10497)

Browse source 
* llama : accept a list of devices to use to offload a model

* accept `--dev none` to completely disable offloading

* fix dev list with dl backends

* rename env parameter to LLAMA_ARG_DEVICE for consistency
This commit is contained in:
Diego Devesa 2024-11-25 19:30:06 +01:00 • committed by GitHub
parent 1f922254f0
commit 10bce0450f
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
9 changed files with 104 additions and 27 deletions
Download patch file Download diff file Expand all files Collapse all files

3
include/llama.h
View file

@ -272,6 +272,9 @@ extern "C" {
};
struct llama_model_params {
// NULL-terminated list of devices to use for offloading (if NULL, all available devices are used)
ggml_backend_dev_t * devices;
int32_t n_gpu_layers; // number of layers to store in VRAM
enum llama_split_mode split_mode; // how to split the model across multiple GPUs