context : do not clear output buffer on reserve (ggml-org#13152)

Co-authored-by: pockers21 <[email protected]>

Author: pockers21

src/llama-context.cpp

@@ -1536,8 +1536,6 @@ int32_t llama_context::output_reserve(int32_t n_outputs) {
     // set all ids as invalid (negative)
     std::fill(output_ids.begin(), output_ids.end(), -1);
 
-    ggml_backend_buffer_clear(buf_output.get(), 0);
-
     this->n_outputs     = 0;
     this->n_outputs_max = n_outputs_max;