Adding fused mul+multi_add + CPU implementation

2026-03-05 03:20:00 +00:00 · 2025-10-23 10:23:59 +03:00
parent db3ba4999f
commit c9b80b2665
1 changed files with 1 additions and 0 deletions
--- a/src/llama-build-context.cpp
+++ b/src/llama-build-context.cpp
@@ -949,6 +949,7 @@ llm_expert_gating_func_type   gating_op,
        }
        experts = ggml_mul(ctx, experts, weights);
        cb(experts, "ffn_moe_weighted", il);
+        return experts;
    }

    if (n_expert_used == 1) {