diff --git a/llama.cpp b/llama.cpp index de51aaf46..f6b8dacdd 100644 --- a/llama.cpp +++ b/llama.cpp @@ -7242,9 +7242,6 @@ struct llm_build_context { cur = moe_out; } - cur = ggml_add(ctx0, cur, ffn_inp); - cb(cur, "ffn_out", il); - // DbrxNormAttentionNorm { cur = llm_build_norm(ctx0, cur, hparams, @@ -7253,6 +7250,9 @@ struct llm_build_context { cb(cur, "layer_out_norm", il); } + cur = ggml_add(ctx0, cur, ffn_inp); + cb(cur, "ffn_out", il); + ggml_tensor * layer_dir = lctx.cvec.tensor_for(il); if (layer_dir != nullptr) { cur = ggml_add(ctx0, cur, layer_dir);