Upload generation_config.json

fix vllm backend startup non-stop generation problem

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 151329,
+  "eos_token_id": 151336,
+  "pad_token_id": 151338,
+  "transformers_version": "4.40.0"
+}