TensorRT同时加载多个模型很容易崩溃

　　比如说，有两个模型同时加载，很容易崩溃。

　　跟踪的时候会发现，在deserialize()的地方出错。

　　如果用gdb看，错误有时是这样的：

[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
Core was generated by `./gh_manager thread false 1'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0  0x00007f7378dfa984 in ?? ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
[Current thread is 1 (Thread 0x7f7345066700 (LWP 18369))]
(gdb) backtrace
#0  0x00007f7378dfa984 in ?? ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
#1  0x00007f7378e83106 in ?? ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
#2  0x00007f7378dff1a7 in ?? ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
#3  0x00007f7378dffe3b in _PyObject_GC_Malloc ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
#4  0x00007f7378dffeed in _PyObject_GC_New ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
#5  0x00007f7378ea43f8 in PyWrapper_New ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
#6  0x00007f7378ea55d7 in ?? ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
#7  0x00007f7378e372b3 in PyObject_Call ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
#8  0x00007f7378dd539c in PyEval_EvalFrameEx ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
#9  0x00007f7378f0e11c in PyEval_EvalCodeEx ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
#10 0x00007f7378e643b0 in ?? ()
   from /usr/lib/x86_64-linux-gnu/libpython2.7.so.1.0
#11 0x00007f7378e372b3 in PyObject_Call ()
---Type  to continue, or q  to quit---q
 from /usr/lib/x86_64-linux-gnu/libpytQuit