From 49138c89f43edc0a70de521080b8a9bf7389ff80 Mon Sep 17 00:00:00 2001
From: LeoTM <1881059+leotm@users.noreply.github.com>
Date: Sat, 13 Jun 2026 20:01:09 +0100
Subject: [PATCH 1/2]  llama : suppress misleading Gemma4Assistant error during
 memory fitting

 Fix: https://github.com/ggml-org/llama.cpp/issues/24343
 Fix: https://github.com/ggml-org/llama.cpp/issues/24350
---
 src/llama-context.cpp | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/src/llama-context.cpp b/src/llama-context.cpp
index 168dbabd7667..b10de6866567 100644
--- a/src/llama-context.cpp
+++ b/src/llama-context.cpp
@@ -22,6 +22,10 @@
 // llama_context
 //
 
+class llama_exception : public std::runtime_error {
+    using std::runtime_error::runtime_error;
+};
+
 static llm_graph_type ctx_type_to_graph_type(llama_context_type ctx_type) {
     switch (ctx_type) {
         case LLAMA_CONTEXT_TYPE_DEFAULT: return LLM_GRAPH_TYPE_DEFAULT;
@@ -93,8 +97,7 @@ llama_context::llama_context(
     // TODO: more generic
     if (model.arch == LLM_ARCH_GEMMA4_ASSISTANT) {
         if (params.ctx_other == nullptr) {
-            // TODO: change from runtime_error to llama_exception to avoid printing error message
-            throw std::runtime_error("Gemma4Assistant requires ctx_other to be set (this warning is normal during memory fitting)");
+            throw llama_exception("Gemma4Assistant requires ctx_other to be set (this warning is normal during memory fitting)");
         }
 
         cparams.ctx_other = params.ctx_other;
@@ -3560,6 +3563,8 @@ llama_context * llama_init_from_model(
     try {
         auto * ctx = new llama_context(*model, params);
         return ctx;
+    } catch (const llama_exception & err) {
+        LLAMA_LOG_WARN("%s: failed to initialize the context: %s\n", __func__, err.what());
     } catch (const std::exception & err) {
         LLAMA_LOG_ERROR("%s: failed to initialize the context: %s\n", __func__, err.what());
     }

From f42196eae454e5667c2a7b666bf129159ec5ff18 Mon Sep 17 00:00:00 2001
From: LeoTM <1881059+leotm@users.noreply.github.com>
Date: Thu, 18 Jun 2026 18:47:46 +0100
Subject: [PATCH 2/2] fixup!  llama : suppress misleading Gemma4Assistant error
 during memory fitting

---
 src/llama-context.cpp | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/llama-context.cpp b/src/llama-context.cpp
index b10de6866567..3559f8fc8206 100644
--- a/src/llama-context.cpp
+++ b/src/llama-context.cpp
@@ -106,7 +106,7 @@ llama_context::llama_context(
     if (model.arch == LLM_ARCH_EAGLE3) {
         if (model.tok_embd == nullptr || model.output == nullptr) {
             if (params.ctx_other == nullptr) {
-                throw std::runtime_error("EAGLE3 requires ctx_other to be set (this warning is normal during memory fitting)");
+                throw llama_exception("EAGLE3 requires ctx_other to be set (this warning is normal during memory fitting)");
             }
             cparams.ctx_other = params.ctx_other;
         }