diff --git a/vllm/entrypoints/scale_out/token_in_token_out/serving.py b/vllm/entrypoints/scale_out/token_in_token_out/serving.py index 631393a43d..90d12f118e 100644 --- a/vllm/entrypoints/scale_out/token_in_token_out/serving.py +++ b/vllm/entrypoints/scale_out/token_in_token_out/serving.py @@ -163,6 +163,8 @@ class ServingTokens(GenerateBaseServing): mm_parser.parse_video(url, uuid) mm_data, mm_uuids = await tracker.resolve_items() prompt = TokensPrompt(prompt_token_ids=request.token_ids) + if request.cache_salt is not None: + prompt["cache_salt"] = request.cache_salt if mm_data: prompt["multi_modal_data"] = mm_data if mm_uuids: