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

Remove mutable variable

Browse source
This commit is contained in:
Yoshi Suhara 2024-08-10 22:28:16 -07:00
parent ae86b5e3d9
commit bd76198618
1 changed files with 0 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

3
src/llama.cpp
View file

@ -13823,9 +13823,6 @@ struct llm_build_context {
struct ggml_cgraph * build_nemotron() { struct ggml_cgraph * build_nemotron() {
struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false); struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
// mutable variable, needed during the last layer of the computation to skip unused tokens
int32_t n_tokens = this->n_tokens;
const int64_t n_embd_head = hparams.n_embd_head_v; const int64_t n_embd_head = hparams.n_embd_head_v;
GGML_ASSERT(n_embd_head == hparams.n_embd_head_k); GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
//GGML_ASSERT(n_embd_head == hparams.n_rot); //GGML_ASSERT(n_embd_head == hparams.n_rot);