root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-25 02:44:36 +00:00
Code Issues Actions 9 Packages Projects Releases Wiki Activity

llama : use n_embd_gqa instead of n_embd to handle llama-2 70B (#2433)

Browse Source
This commit is contained in:
Rand Xie 2023-07-28 01:42:53 -07:00 committed by GitHub
parent edcc7ae7d2
commit 65cdf34bdc
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 3 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
examples/save-load-state/save-load-state.cpp
View File

@ -26,6 +26,7 @@ int main(int argc, char ** argv) {
auto lparams = llama_context_default_params(); auto lparams = llama_context_default_params();
lparams.n_ctx = params.n_ctx; lparams.n_ctx = params.n_ctx;
lparams.n_gqa = params.n_gqa;
lparams.seed = params.seed; lparams.seed = params.seed;
lparams.f16_kv = params.memory_f16; lparams.f16_kv = params.memory_f16;
lparams.use_mmap = params.use_mmap; lparams.use_mmap = params.use_mmap;

4
llama.cpp
View File

@ -3663,7 +3663,7 @@ size_t llama_copy_state_data(struct llama_context * ctx, uint8_t * dst) {
const auto & kv_self = ctx->kv_self; const auto & kv_self = ctx->kv_self;
const auto & hparams = ctx->model.hparams; const auto & hparams = ctx->model.hparams;
const int n_layer = hparams.n_layer; const int n_layer = hparams.n_layer;
const int n_embd = hparams.n_embd; const int n_embd = hparams.n_embd_gqa();
const int n_ctx = hparams.n_ctx; const int n_ctx = hparams.n_ctx;
const size_t kv_size = kv_self.buf.size; const size_t kv_size = kv_self.buf.size;
@ -3766,7 +3766,7 @@ size_t llama_set_state_data(struct llama_context * ctx, uint8_t * src) {
const auto & kv_self = ctx->kv_self; const auto & kv_self = ctx->kv_self;
const auto & hparams = ctx->model.hparams; const auto & hparams = ctx->model.hparams;
const int n_layer = hparams.n_layer; const int n_layer = hparams.n_layer;
const int n_embd = hparams.n_embd; const int n_embd = hparams.n_embd_gqa();
const int n_ctx = hparams.n_ctx; const int n_ctx = hparams.n_ctx;
size_t kv_size; size_t kv_size;