models : move the token embedding norms to the first layer (#20943)

* models : move the token embedding norms to the first layer * cont : fix LLM_TENSOR_CONV1D + fix il indexing
2026-03-24 17:00:30 +02:00
parent 3fc6f1aed1
commit 9f102a1407
8 changed files with 26 additions and 26 deletions
@@ -93,7 +93,7 @@ llm_build_wavtokenizer_dec::llm_build_wavtokenizer_dec(const llama_model & model
    cur = build_norm(cur,
            model.tok_norm,
            model.tok_norm_b,
-            LLM_NORM, -1);
+            LLM_NORM, 0);

    cur = ggml_cont(ctx0, ggml_transpose(ctx0, cur));