[BUG] [DLLM] Missing max_running_requests value (#18740)
This commit is contained in:
@@ -60,7 +60,7 @@ class SchedulerDllmMixin:
|
|||||||
|
|
||||||
def _fetch_waiting_reqs(self: Scheduler):
|
def _fetch_waiting_reqs(self: Scheduler):
|
||||||
# Calculate how many requests can be added to DLLM manager
|
# Calculate how many requests can be added to DLLM manager
|
||||||
max_dllm_capacity = self.server_args.max_running_requests - len(
|
max_dllm_capacity = self.dllm_config.max_running_requests - len(
|
||||||
self.dllm_manager.waiting_queue
|
self.dllm_manager.waiting_queue
|
||||||
)
|
)
|
||||||
num_requests_to_add = min(max_dllm_capacity, len(self.waiting_queue))
|
num_requests_to_add = min(max_dllm_capacity, len(self.waiting_queue))
|
||||||
|
|||||||
Reference in New Issue
Block a user