diff --git a/examples/quantize/test.sh b/examples/quantize/test.sh
index fe64f09757d37..840f712accfda 100644
--- a/examples/quantize/test.sh
+++ b/examples/quantize/test.sh
@@ -42,7 +42,7 @@ echo PASS
 echo
 
 # 3. Requant model with '--keep_split'
-$QUANTIZE --allow-requantize --keep_split $WORK_PATH/ggml-model-split-00001-of-00006.gguf $WORK_PATH/ggml-model-requant.gguf Q4_K 
+$QUANTIZE --allow-requantize --keep_split $WORK_PATH/ggml-model-split-00001-of-00006.gguf $WORK_PATH/ggml-model-requant.gguf Q4_K
 echo PASS
 echo
 
@@ -52,7 +52,7 @@ echo PASS
 echo
 
 # 4. Requant mode without '--keep_split'
-$QUANTIZE --allow-requantize $WORK_PATH/ggml-model-split-00001-of-00006.gguf $WORK_PATH/ggml-model-requant-merge.gguf Q4_K 
+$QUANTIZE --allow-requantize $WORK_PATH/ggml-model-split-00001-of-00006.gguf $WORK_PATH/ggml-model-requant-merge.gguf Q4_K
 echo PASS
 echo
 
@@ -62,4 +62,4 @@ echo PASS
 echo
 
 # Clean up
-rm -f $WORK_PATH/ggml-model-split*.gguf $WORK_PATH/ggml-model-requant*.gguf
\ No newline at end of file
+rm -f $WORK_PATH/ggml-model-split*.gguf $WORK_PATH/ggml-model-requant*.gguf
diff --git a/llama.cpp b/llama.cpp
index 550fc67351adc..a7f5da88dddf8 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -13618,13 +13618,13 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s
 
     // Set split info if needed
     if (n_split > 1) {
-        for (int i = 0; i < ctx_outs.size(); ++i) {
+        for (size_t i = 0; i < ctx_outs.size(); ++i) {
             gguf_set_val_u16(ctx_outs[i], ml.llm_kv(LLM_KV_SPLIT_NO).c_str(), i);
             gguf_set_val_u16(ctx_outs[i], ml.llm_kv(LLM_KV_SPLIT_COUNT).c_str(), n_split);
             gguf_set_val_i32(ctx_outs[i], ml.llm_kv(LLM_KV_SPLIT_TENSORS_COUNT).c_str(), ml.n_tensors);
         }
     }
-    
+
     int cur_split = -1;
     std::ofstream fout;
     auto close_ofstream = [&]() {