Skip to content

LLM inference in C/C++

C++ ggml

model: Add Laguna-S-2.1 LLM_TYPE (#26233)

Guido Imperialee9fa0781f1c2
diff --git a/src/llama-model.cpp b/src/llama-model.cpp
index be0a0df55d..7a70585fa4 100644
--- a/src/llama-model.cpp
+++ b/src/llama-model.cpp
@@ -818,6 +818,7 @@ const char * llm_type_name(llm_type type) {
case LLM_TYPE_100B_A6B: return "100B.A6B";
case LLM_TYPE_102B_A12B: return "102B.A12B";
case LLM_TYPE_106B_A12B: return "106B.A12B";
+ case LLM_TYPE_118B_A8B: return "118B.A8B";
case LLM_TYPE_120B_A12B: return "120B.A12B";
case LLM_TYPE_122B_A10B: return "122B.A10B";
case LLM_TYPE_196B_A11B: return "196B.A11B";
diff --git a/src/llama-model.h b/src/llama-model.h
index d6a40fa302..056a6efa59 100644
--- a/src/llama-model.h
+++ b/src/llama-model.h
@@ -130,6 +130,7 @@ enum llm_type {
LLM_TYPE_100B_A6B,
LLM_TYPE_102B_A12B, // Solar-Open
LLM_TYPE_106B_A12B, // GLM-4.5-Air
+ LLM_TYPE_118B_A8B, // Laguna-S-2
LLM_TYPE_120B_A12B, // Nemotron 3 Super
LLM_TYPE_122B_A10B, // Qwen3.5
LLM_TYPE_196B_A11B, // Step3.5-Flash
diff --git a/src/models/laguna.cpp b/src/models/laguna.cpp
index fb55ec12f9..82c9a9538c 100644
--- a/src/models/laguna.cpp
+++ b/src/models/laguna.cpp
@@ -58,6 +58,7 @@ void llama_model_laguna::load_arch_hparams(llama_model_loader & ml) {
switch (hparams.n_layer()) {
case 40: type = LLM_TYPE_30B_A3B; break; // Laguna-XS.2
+ case 48: type = LLM_TYPE_118B_A8B; break; // Laguna-S.2
case 70: type = LLM_TYPE_230B_A10B; break; // Laguna-M.1
default: type = LLM_TYPE_UNKNOWN;
}