[ModelRunner V2] Enable by default for all dense models (#44443)
Signed-off-by: yewentao256 <zhyanwentao@126.com> Signed-off-by: Nick Hill <nickhill123@gmail.com> Co-authored-by: Nick Hill <nickhill123@gmail.com>
This commit is contained in:
@@ -84,7 +84,13 @@ def assert_executor(executor, tp_size, pp_size):
|
||||
assert executor._get_output_rank() == expected_output_rank
|
||||
|
||||
if pp_size > 1:
|
||||
assert executor.vllm_config.max_concurrent_batches == pp_size
|
||||
expected_concurrent_batches = pp_size + int(
|
||||
executor.vllm_config.scheduler_config.async_scheduling
|
||||
and executor.vllm_config.use_v2_model_runner
|
||||
)
|
||||
assert (
|
||||
executor.vllm_config.max_concurrent_batches == expected_concurrent_batches
|
||||
)
|
||||
|
||||
executor.check_health()
|
||||
assert not executor.is_failed
|
||||
|
||||
Reference in New Issue
Block a user