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

llama : fix not enough space in buffer with Qwen

Browse source
This commit is contained in:
slaren 2024-01-22 21:05:25 +01:00
parent 6f9939d119
commit f0bb1052c6
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

@ -4440,9 +4440,9 @@ static struct ggml_tensor * llm_build_kv(
// these nodes are added to the graph together so that they are not reordered
// by doing so, the number of splits in the graph is reduced
ggml_build_forward_expand(graph, q_cur);
ggml_build_forward_expand(graph, k_cur);
ggml_build_forward_expand(graph, v_cur);
ggml_build_forward_expand(graph, q_cur);
llm_build_kv_store(ctx, hparams, kv, graph, k_cur, v_cur, n_ctx, n_tokens, kv_head, cb, il);