We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent f82f050 commit e5c32f2Copy full SHA for e5c32f2
src/llama-graph.cpp
@@ -929,7 +929,7 @@ ggml_tensor * llm_graph_context::build_moe_ffn(
929
}
930
931
// select top n_group_used expert groups
932
- if (arch == LLM_ARCH_BAILINGMOE2) {
+ if (arch == LLM_ARCH_BAILINGMOE2 && n_tokens > 0) {
933
const int64_t n_exp_per_group = n_expert / hparams.n_expert_groups;
934
935
// organize experts into n_expert_groups
0 commit comments