router(grpc): Implement route for chat_cmpl endpoint (#10761)

2025-09-23 11:26:33 -07:00
parent b24b2e7ed7
commit 7ff93e613f
8 changed files with 380 additions and 11 deletions
--- a/sgl-router/src/proto/sglang_scheduler.proto
+++ b/sgl-router/src/proto/sglang_scheduler.proto
@@ -36,7 +36,7 @@ message SamplingParams {
  float presence_penalty = 6;
  float repetition_penalty = 7;

-  int32 max_new_tokens = 8;
+  optional int32 max_new_tokens = 8;
  repeated string stop = 9;
  repeated int32 stop_token_ids = 10;
  bool skip_special_tokens = 11;