Created by: barrierye

python端client跨线程用numpy预测的时候会偶发性报错（在不报错的情况下预测结果正常），list暂时没遇到这种情况。 可能是GIL的释放导致变量生命周期管理异常（list对应的predict的GIL部分不知道什么时候不见了）

报错信息：

# python profile_org_client.py
WARNING: Logging before InitGoogleLogging() is written to STDERR
I0522 18:25:04.466261 28517 naming_service_thread.cpp:209] brpc::policy::ListNamingService("127.0.0.1:9393"): added 1
Exception in thread Thread-3:
Traceback (most recent call last):
  File "/usr/lib64/python2.7/threading.py", line 812, in __bootstrap_inner
    self.run()
  File "/usr/lib64/python2.7/threading.py", line 765, in run
    self.__target(*self.__args, **self.__kwargs)
  File "profile_org_client.py", line 37, in func
    feed={"x": x}, fetch=["price"])
  File "/home/barriery/.local/lib/python2.7/site-packages/paddle_serving_client/__init__.py", line 342, in predict
    result_map[name].shape = shape
ValueError: cannot reshape array of size 0 into shape (1,1)

Exception in thread Thread-2:
Traceback (most recent call last):
  File "/usr/lib64/python2.7/threading.py", line 812, in __bootstrap_inner
    self.run()
  File "/usr/lib64/python2.7/threading.py", line 765, in run
    self.__target(*self.__args, **self.__kwargs)
  File "profile_org_client.py", line 37, in func
    feed={"x": x}, fetch=["price"])
  File "/home/barriery/.local/lib/python2.7/site-packages/paddle_serving_client/__init__.py", line 342, in predict
    result_map[name].shape = shape
ValueError: cannot reshape array of size 0 into shape (1,1)

复现代码：

from paddle_serving_client import Client
import numpy as np
import sys
import threading
import time

from line_profiler import LineProfiler
client = Client()
client.load_client_config("uci_housing_client/serving_client_conf.prototxt")
client.connect(["127.0.0.1:9393"])

import paddle

lx = [0.0137, -0.1136, 0.2553, -0.0692, 0.0582, -0.0727, -0.1583, -0.0584,
        0.6283, 0.4919, 0.1856, 0.0795, -0.0332]
x = np.array(lx, dtype='float')

def func(num):

    for i in range(num):
        fetch_map = client.predict(
            feed={"x": x}, fetch=["price"])
        # print(fetch_map)

thread_num = 3
each_thread = 1000
ths = []
start = time.time()

for i in range(thread_num):
    th = threading.Thread(target=func, args=(each_thread, ))
    th.start()
    ths.append(th)

for i in range(thread_num):
    ths[i].join()

end = time.time()
print("time: {}".format(end - start))