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

code : normalize enum names (#5697)

Browse source 
* coda : normalize enum names

ggml-ci

* code : cont

* code : cont
This commit is contained in:
Georgi Gerganov 2024-02-25 12:09:09 +02:00 • committed by GitHub
parent 69917dfa55
commit ab336a9d5e
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
20 changed files with 502 additions and 502 deletions
Download patch file Download diff file Expand all files Collapse all files

4
common/common.h
View file

@ -61,7 +61,7 @@ struct gpt_params {
float p_split = 0.1f; // speculative decoding split probability
int32_t n_gpu_layers = -1; // number of layers to store in VRAM (-1 - use default)
int32_t n_gpu_layers_draft = -1; // number of layers to store in VRAM for the draft model (-1 - use default)
llama_split_mode split_mode = LLAMA_SPLIT_LAYER; // how to split the model across GPUs
llama_split_mode split_mode = LLAMA_SPLIT_MODE_LAYER; // how to split the model across GPUs
int32_t main_gpu = 0; // the GPU that is used for scratch and small tensors
float tensor_split[128] = {0}; // how split tensors should be distributed across GPUs
int32_t n_beams = 0; // if non-zero then use beam search of given width.
@ -75,7 +75,7 @@ struct gpt_params {
float yarn_beta_fast = 32.0f; // YaRN low correction dim
float yarn_beta_slow = 1.0f; // YaRN high correction dim
int32_t yarn_orig_ctx = 0; // YaRN original context length
int32_t rope_scaling_type = LLAMA_ROPE_SCALING_UNSPECIFIED;
int32_t rope_scaling_type = LLAMA_ROPE_SCALING_TYPE_UNSPECIFIED;
ggml_numa_strategy numa = GGML_NUMA_STRATEGY_DISABLED;
// // sampling parameters