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

rename n_ctx to kv_size

Browse source
This commit is contained in:
Pierrick HYMBERT 2024-02-18 20:59:26 +01:00 • committed by Georgi Gerganov
parent ef96e8b1f7
commit 606873401c
No known key found for this signature in database
GPG key ID: 449E073F9DC10735
48 changed files with 403 additions and 393 deletions
Download patch file Download diff file Expand all files Collapse all files

2
common/common.h
View file

@ -50,7 +50,7 @@ struct gpt_params {
int32_t n_threads_batch = -1; // number of threads to use for batch processing (-1 = use n_threads)
int32_t n_threads_batch_draft = -1;
int32_t n_predict = -1; // new tokens to predict
int32_t n_ctx = 512; // context size
int32_t kv_size = 512; // KV Cache size
int32_t n_batch = 512; // batch size for prompt processing (must be >=32 to use BLAS)
int32_t n_keep = 0; // number of tokens to keep from initial prompt
int32_t n_draft = 8; // number of tokens to draft during speculative decoding