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

No need to copy tokens

Browse source
This commit is contained in:
Howard Su 2023-04-17 23:07:30 +08:00
parent efd05648c8
commit 48ab0963ae
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
llama.cpp
View file

@ -1066,7 +1066,7 @@ static bool llama_eval_internal(
gf.n_threads = N >= 32 && ggml_cpu_has_blas() ? 1 : n_threads;
struct ggml_tensor * embd = ggml_new_tensor_1d(ctx0, GGML_TYPE_I32, N);
memcpy(embd->data, tokens, N*ggml_element_size(embd));
embd->data = (void*)tokens;
struct ggml_tensor * inpL = ggml_get_rows(ctx0, model.tok_embeddings, embd);