diff --git a/examples/llava/clip.cpp b/examples/llava/clip.cpp
index 74c961fc4..ec0a49611 100644
--- a/examples/llava/clip.cpp
+++ b/examples/llava/clip.cpp
@@ -2177,7 +2177,7 @@ bool clip_image_preprocess(struct clip_ctx * ctx, const clip_image_u8 * img, cli
         return true;
     }
 
-    if(ctx->has_glm_projector){
+    if (ctx->has_glm_projector) {
         res_imgs->size = 1;
         res_imgs->data = new clip_image_f32[res_imgs->size];
         clip_image_u8 resized_image;
@@ -2376,9 +2376,8 @@ void clip_free(clip_ctx * ctx) {
 }
 
 size_t clip_embd_nbytes(const struct clip_ctx * ctx) {
-    if(ctx->has_glm_projector)
-        return (clip_n_patches(ctx)+2) * clip_n_mmproj_embd(ctx) * sizeof(float);
-    return clip_n_patches(ctx) * clip_n_mmproj_embd(ctx) * sizeof(float);
+    int extra_tokens = ctx->has_glm_projector ? 2 : 0;
+    return (clip_n_patches(ctx) + extra_tokens) * clip_n_mmproj_embd(ctx) * sizeof(float);
 }
 
 size_t clip_embd_nbytes_by_img(const struct clip_ctx * ctx, int img_h, int img_w) {
@@ -2553,11 +2552,11 @@ bool clip_image_batch_encode(clip_ctx * ctx, const int n_threads, const clip_ima
     if (ctx->has_minicpmv_projector) {
         GGML_ASSERT(batch_size == 1);
     }
-    if(ctx->has_glm_projector) {
+    if (ctx->has_glm_projector) {
         GGML_ASSERT(batch_size == 1);
         ggml_tensor * boi = ctx->vision_model.boi_w;
         ggml_backend_tensor_get(boi,vec,0,ggml_nbytes(boi));
-        vec=(float*)(vec+ggml_nelements(boi)); //offset for boi
+        vec = (float*)(vec+ggml_nelements(boi)); //offset for boi
     }
 
     // build the inference graph
@@ -2711,7 +2710,7 @@ bool clip_image_batch_encode(clip_ctx * ctx, const int n_threads, const clip_ima
             ggml_backend_tensor_set(positions, positions_data, 0, ggml_nbytes(positions));
             free(positions_data);
 
-            if (!ctx->has_glm_projector){
+            if (!ctx->has_glm_projector) {
                 struct ggml_tensor * patches = ggml_graph_get_tensor(gf, "patches");
                 int* patches_data = (int*)malloc(ggml_nbytes(patches));
                 for (int i = 0; i < num_patches; i++) {
@@ -2735,11 +2734,11 @@ bool clip_image_batch_encode(clip_ctx * ctx, const int n_threads, const clip_ima
     // copy the embeddings to the location passed by the user
     ggml_backend_tensor_get(embeddings, vec, 0, ggml_nbytes(embeddings));
 
-    if(ctx->has_glm_projector){
+    if (ctx->has_glm_projector) {
         //eoi
         ggml_tensor * eoi = ctx->vision_model.eoi_w;
         int offset = ggml_nelements(embeddings);
-        ggml_backend_tensor_get(eoi,vec+offset,0,ggml_nbytes(eoi));
+        ggml_backend_tensor_get(eoi, vec+offset, 0, ggml_nbytes(eoi));
     }
 
     return true;
diff --git a/src/llama-model.cpp b/src/llama-model.cpp
index ad0e96b5a..fdaffba80 100644
--- a/src/llama-model.cpp
+++ b/src/llama-model.cpp
@@ -1093,20 +1093,20 @@ void llama_model::load_hparams(llama_model_loader & ml) {
             {
                 ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
                 switch (hparams.n_layer) {
-                    case 28:{
-                        if(hparams.n_head(0)==16){
+                    case 28: {
+                        if (hparams.n_head(0)==16) {
                             type = LLM_TYPE_1_5B;
-                        }else{
+                        } else {
                             type = LLM_TYPE_6B;
                         }
                     } break;
                     case 40: {
                         if(hparams.n_head(0)==24){
                             type = LLM_TYPE_4B;
-                        }else{
+                        } else {
                             type = LLM_TYPE_9B;
                         }
-                    }break;
+                    } break;
                     default: type = LLM_TYPE_UNKNOWN;
                 }
             } break;
diff --git a/src/llama.cpp b/src/llama.cpp
index a77d9d7ae..6c4e59b25 100644
--- a/src/llama.cpp
+++ b/src/llama.cpp
@@ -7215,7 +7215,7 @@ struct llm_build_context {
                 struct ggml_tensor * Qcur = nullptr;
                 struct ggml_tensor * Kcur = nullptr;
                 struct ggml_tensor * Vcur = nullptr;
-                if(model.type == LLM_TYPE_1_5B|| model.type == LLM_TYPE_4B || model.type == LLM_TYPE_9B) {
+                if (model.type == LLM_TYPE_1_5B || model.type == LLM_TYPE_4B || model.type == LLM_TYPE_9B) {
                     Qcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wq, cur);
                     cb(Qcur, "Qcur", il);
                     if (model.layers[il].bq) {
@@ -7234,10 +7234,10 @@ struct llm_build_context {
                         Vcur = ggml_add(ctx0, Vcur, model.layers[il].bv);
                         cb(Vcur, "Vcur", il);
                     }
-                }else{
+                } else {
                     cur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wqkv, cur);
                     cb(cur, "wqkv", il);
-                    if(model.layers[il].bqkv){
+                    if (model.layers[il].bqkv) {
                         cur = ggml_add(ctx0, cur, model.layers[il].bqkv);
                         cb(cur, "bqkv", il);
                     }