sequence batcher coredump when client not send end flag

[wuwenjunwwj]

Description
when I deploy funasr streaming models with triton server，when client not send end flag，the triton server coredump 。https://github.com/modelscope/FunASR/tree/main/runtime/triton_gpu/model_repo_paraformer_large_online

the coredump is：
#0 0x00007f41048bf4d1 in triton::core::InferenceRequest::LogRequestabi:cxx11 const () from /opt/tritonserver/lib/libtritonserver.so
[Current thread is 1 (Thread 0x7f3a17fff000 (LWP 30710))]
(gdb) bt
#0 0x00007f41048bf4d1 in triton::core::InferenceRequest::LogRequestabi:cxx11 const () from /opt/tritonserver/lib/libtritonserver.so
#1 0x00007f41049f4ecf in triton::core::OldestSequenceBatch::CompleteAndNext(unsigned int) () from /opt/tritonserver/lib/libtritonserver.so
#2 0x00007f41049f370b in triton::core::SequenceBatchScheduler::ReaperThread(int) () from /opt/tritonserver/lib/libtritonserver.so
#3 0x00007f4104403de4 in ?? () from /usr/lib/x86_64-linux-gnu/libstdc++.so.6
#4 0x00007f410c8aa609 in start_thread (arg=) at pthread_create.c:477
#5 0x00007f41040f0353 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95

Triton Information
What version of Triton are you using?
r23.04

Are you using the Triton container or did you build it yourself?

To Reproduce

Expected behavior
A clear and concise description of what you expected to happen.