From 89eaf5036a07b394bea4c7b35f97d171bee2e87b Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Mon, 16 Dec 2024 21:03:24 +0200
Subject: [PATCH] server : add "tokens" output

ggml-ci
---
 examples/server/server.cpp | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/examples/server/server.cpp b/examples/server/server.cpp
index 5ed4e8d27..57db582d7 100644
--- a/examples/server/server.cpp
+++ b/examples/server/server.cpp
@@ -687,8 +687,6 @@ struct server_task_result_cmpl_partial : server_task_result {
                 json second_ret = json{
                             {"choices", json::array({json{{"finish_reason", nullptr},
                                                             {"index", 0},
-                                                            {"delta", json {
-                                                            {"content", content}}}
                                                             }})},
                             {"created", t},
                             {"id", oaicompat_cmpl_id},
@@ -704,6 +702,7 @@ struct server_task_result_cmpl_partial : server_task_result {
                 {"delta",
                 json {
                     {"content", content},
+                    {"tokens",  tokens}
                 }},
             }});
         }
@@ -1017,6 +1016,7 @@ struct server_slot {
         n_prompt_tokens    = 0;
         last_nl_pos        = 0;
         generated_text     = "";
+        generated_tokens   = {};
         has_new_line       = false;
         truncated          = false;
         stop               = STOP_TYPE_NONE;