[Core] Raise when non-multi-instance DP clients target a DP rank (#19227)
Signed-off-by: Jon Swenson <jmswen@gmail.com>
This commit is contained in:
@@ -494,6 +494,10 @@ class _AsyncLLMEngine(LLMEngine):
|
||||
if arrival_time is None:
|
||||
arrival_time = time.time()
|
||||
|
||||
if data_parallel_rank is not None:
|
||||
raise ValueError("Targeting data_parallel_rank only supported "
|
||||
"in v1 client.")
|
||||
|
||||
if (isinstance(prompt, dict)
|
||||
and prompt.get("prompt_embeds", None) is not None
|
||||
and not prompt.get("prompt_token_ids", None)):
|
||||
|
||||
Reference in New Issue
Block a user