Update deps for mllama4 (#5215)

This commit is contained in:
Ke Bao
2025-04-11 00:12:44 +08:00
committed by GitHub
parent 7e4f72dd8c
commit 1078396f47
3 changed files with 26 additions and 25 deletions

View File

@@ -840,7 +840,6 @@ class Scheduler(
bootstrap_room=recv_req.bootstrap_room,
)
req.tokenizer = self.tokenizer
req.queue_time_start = time.time()
if (
recv_req.session_params is not None
@@ -855,7 +854,6 @@ class Scheduler(
# Create a new request from a previous session
session = self.sessions[recv_req.session_params.id]
req = session.create_req(recv_req, self.tokenizer)
req.queue_time_start = time.time()
if isinstance(req.finished_reason, FINISH_ABORT):
self._add_request_to_queue(req)
return
@@ -958,6 +956,7 @@ class Scheduler(
self.disagg_decode_prealloc_queue.add(req)
else:
req.queue_time_start = time.time()
self.waiting_queue.append(req)
def _extend_requests_to_queue(self, reqs: List[Req], is_retracted: bool = False):