ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

llama : grouped-query attention + LLaMAv2 70B support (#2276)

Browse source 
* CUDA: GQA implementation

* llama : support for GQA and LLaMAv2 70B

ggml-ci

* py : fix hparams parsing (if-else blocks)

ggml-ci

* py : oh boy ..

ggml-ci

* help : fix gqa value for 70B

ggml-ci

---------

Co-authored-by: JohannesGaessler <johannesg@5d6.de>
This commit is contained in:
Georgi Gerganov 2023-07-23 15:09:47 +03:00 • committed by GitHub
parent 1d0824b247
commit e76d630df1
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
7 changed files with 215 additions and 108 deletions
Download patch file Download diff file Expand all files Collapse all files

3
examples/common.h
View file

@ -27,6 +27,7 @@ struct gpt_params {
int32_t n_predict = -1; // new tokens to predict
int32_t n_ctx = 512; // context size
int32_t n_batch = 512; // batch size for prompt processing (must be >=32 to use BLAS)
int32_t n_gqa = 1; // grouped-query attention factor (TODO: move to hparams)
int32_t n_keep = 0; // number of tokens to keep from initial prompt
int32_t n_chunks = -1; // max number of chunks to process (-1 = unlimited)
int32_t n_gpu_layers = 0; // number of layers to store in VRAM
@ -47,7 +48,7 @@ struct gpt_params {
int32_t repeat_last_n = 64; // last n tokens to penalize (0 = disable penalty, -1 = context size)
float frequency_penalty = 0.00f; // 0.0 = disabled
float presence_penalty = 0.00f; // 0.0 = disabled
int mirostat = 0; // 0 = disabled, 1 = mirostat, 2 = mirostat 2.0
int32_t mirostat = 0; // 0 = disabled, 1 = mirostat, 2 = mirostat 2.0
float mirostat_tau = 5.00f; // target entropy
float mirostat_eta = 0.10f; // learning rate