iAdding support for dense Qwen-3.5 models (#1326)

src/llama.cpp

@@ -5358,6 +5358,7 @@ enum llama_rope_type llama_rope_type(const struct llama_model * model) {
         case LLM_ARCH_QWEN3VL:
         case LLM_ARCH_QWEN3VLMOE:
         case LLM_ARCH_QWEN35MOE:
+        case LLM_ARCH_QWEN35:
             return LLAMA_ROPE_TYPE_IMROPE;
 
         // all model arches should be listed explicitly here