From 5a4fa17947e4df1a7fb0afa06a20276e1065b92c Mon Sep 17 00:00:00 2001 From: Yap Sok Ann Date: Wed, 24 Jun 2026 14:04:09 +0700 Subject: [PATCH] Load glm-dsa indexer tensors as optional (ggml-org/llama.cpp#24770) (#2017) --- src/llama-load-tensors.cpp | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) diff --git a/src/llama-load-tensors.cpp b/src/llama-load-tensors.cpp index ccc2962b..33ca9553 100644 --- a/src/llama-load-tensors.cpp +++ b/src/llama-load-tensors.cpp @@ -2788,11 +2788,15 @@ bool create_tensors_helper::create_glm_dsa_tensors(const LLM_TN & tn) { layer.wo = create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_head * n_embd_head_v, n_embd}, flags); // DSA indexer - layer.indexer_k_norm = create_tensor(ctx_split, tn(LLM_TENSOR_INDEXER_K_NORM, "weight", i), {hparams.indexer_head_size}, flags); - layer.indexer_k_norm_b = create_tensor(ctx_split, tn(LLM_TENSOR_INDEXER_K_NORM, "bias", i), {hparams.indexer_head_size}, flags); - layer.indexer_proj = create_tensor(ctx_split, tn(LLM_TENSOR_INDEXER_PROJ, "weight", i), {n_embd, hparams.indexer_n_head}, flags); - layer.indexer_attn_k = create_tensor(ctx_split, tn(LLM_TENSOR_INDEXER_ATTN_K, "weight", i), {n_embd, hparams.indexer_head_size}, flags); - layer.indexer_attn_q_b = create_tensor(ctx_split, tn(LLM_TENSOR_INDEXER_ATTN_Q_B, "weight", i), {q_lora_rank, hparams.indexer_n_head * hparams.indexer_head_size}, flags); + // GLM-5.2 only ships the DSA indexer on a subset of layers; the rest omit it. + // The DSA indexer runtime is not implemented (graph is plain MLA), so these + // tensors are loaded-but-unused. Mark them optional so layers without an + // indexer load as nullptr (ported from ggml-org/llama.cpp#24770). + layer.indexer_k_norm = create_tensor(ctx_split, tn(LLM_TENSOR_INDEXER_K_NORM, "weight", i), {hparams.indexer_head_size}, flags | llama_model_loader::TENSOR_NOT_REQUIRED); + layer.indexer_k_norm_b = create_tensor(ctx_split, tn(LLM_TENSOR_INDEXER_K_NORM, "bias", i), {hparams.indexer_head_size}, flags | llama_model_loader::TENSOR_NOT_REQUIRED); + layer.indexer_proj = create_tensor(ctx_split, tn(LLM_TENSOR_INDEXER_PROJ, "weight", i), {n_embd, hparams.indexer_n_head}, flags | llama_model_loader::TENSOR_NOT_REQUIRED); + layer.indexer_attn_k = create_tensor(ctx_split, tn(LLM_TENSOR_INDEXER_ATTN_K, "weight", i), {n_embd, hparams.indexer_head_size}, flags | llama_model_loader::TENSOR_NOT_REQUIRED); + layer.indexer_attn_q_b = create_tensor(ctx_split, tn(LLM_TENSOR_INDEXER_ATTN_Q_B, "weight", i), {q_lora_rank, hparams.indexer_n_head * hparams.indexer_head_size}, flags | llama_model_loader::TENSOR_NOT_REQUIRED); layer.ffn_norm = create_tensor(norm_ctx, tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, flags);