Fix spec info's filter when reqs are finished right after prefill (#14742)

python/sglang/srt/managers/scheduler.py

@@ -1937,7 +1937,7 @@ class Scheduler(
             and not (new_batch.return_logprob or self.running_batch.return_logprob)
         ):
             # TODO (lianmin): support return_logprob + mixed chunked prefill
-            self.running_batch.filter_batch()
+            self.running_batch.filter_batch(v1_spec_info_filtered=True)
             if not self.running_batch.is_empty():
                 self.running_batch.prepare_for_decode()
                 new_batch.mix_with_running(self.running_batch)
@@ -1954,7 +1954,7 @@ class Scheduler(
         """Update the current running decoding batch."""
         initial_bs = batch.batch_size()
 
-        batch.filter_batch()
+        batch.filter_batch(v1_spec_info_filtered=True)
         if batch.is_empty():
             batch.batch_is_full = False
             return batch
@@ -2509,7 +2509,7 @@ class Scheduler(
         self.cur_batch = None
 
         if recv_req.mode == "retract":
-            self.running_batch.filter_batch()
+            self.running_batch.filter_batch(v1_spec_info_filtered=True)
             if len(self.running_batch.reqs) != 0:
                 retracted_reqs = self.running_batch.retract_all(self.server_args)
                 for req in retracted_reqs: