From 686fd1ebec8a64f7e8221d5efa6363a699aa19a1 Mon Sep 17 00:00:00 2001 From: Kawrakow Date: Thu, 29 Jan 2026 09:06:40 +0200 Subject: [PATCH] Use standard output calculation for MiniMax-M2 graph parallel (#1199) --- src/llama-build-context.cpp | 12 +----------- 1 file changed, 1 insertion(+), 11 deletions(-) diff --git a/src/llama-build-context.cpp b/src/llama-build-context.cpp index 7da9a4e1..e10e6cfe 100644 --- a/src/llama-build-context.cpp +++ b/src/llama-build-context.cpp @@ -8833,17 +8833,7 @@ ggml_cgraph* llm_build_context::build_minimaxm2() { inpL = cur; } - cur = inpL; - - cur = llm_build_norm(ctx0, cur, - hparams, model.output_norm, NULL, - LLM_NORM_RMS, cb, -1); - - cb(cur, "result_norm", -1); - - // lm_head - cur = llm_build_lora_mm(lctx, ctx0, model.output, cur); - + cur = build_output(lctx, ctx0, inpL, model.output, model.output_norm, cb); cb(cur, "result_output", -1); ggml_build_forward_expand(gf, cur);