Bugfix: Ensure logs are printed when streaming

Author: abetlen

llama_cpp/llama.py

@@ -709,6 +709,9 @@ def _create_completion(
                 print("Llama._create_completion: cache save", file=sys.stderr)
             self.cache[prompt_tokens + completion_tokens] = self.save_state()
 
+        if self.verbose:
+            llama_cpp.llama_print_timings(self.ctx)
+
         if stream:
             yield {
                 "id": completion_id,
@@ -780,9 +783,6 @@ def _create_completion(
                 "top_logprobs": top_logprobs,
             }
 
-        if self.verbose:
-            llama_cpp.llama_print_timings(self.ctx)
-
         yield {
             "id": completion_id,
             "object": "text_completion",