We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent d0b8fd3 commit 05d8b6aCopy full SHA for 05d8b6a
vllm_ascend/worker/worker_v1.py
@@ -277,9 +277,10 @@ def pin_lora(self, lora_id: int) -> bool:
277
278
def execute_dummy_batch(self) -> None:
279
runner = self.model_runner
280
- assert runner.dp_size > 1, "Dummy batch execution should only be " \
281
- "performed with data parallelism enabled, but got " \
282
- f"dp_size={runner.dp_size}."
+ if runner.dp_size <= 1:
+ raise ValueError("Dummy batch execution should only be "
+ "performed with data parallelism enabled, but got "
283
+ f"dp_size={runner.dp_size}.")
284
285
# If torchair graph is enabled, notify the other DP ranks that this is a
286
# dummy run by using '-1' as a flag for num_tokens. This will be
0 commit comments