2021-10-12 Paddle DeepSpeech生成数据列表报错

执行python create_data.py生成数据列表过程中报错：

开始抽取-1条数据计算均值和标准值...
  0%|                                                                                                                                                             | 0/4017 [00:00 W1011 18:34:43.968842 56947 naive_best_fit_allocator.cc:519] cudaHostAlloc Cannot allocate 1288 bytes in CUDAPinnedPlace
 

--------------------------------------
C++ Traceback (most recent call last):
--------------------------------------
0   std::thread::_State_impl > >::_M_run()
1   std::__future_base::_State_baseV2::_M_do_set(std::function ()>*, bool*)
2   paddle::framework::SignalHandle(char const*, int)
3   paddle::platform::GetCurrentTraceBackString[abi:cxx11]()

----------------------
Error Message Summary:
----------------------
FatalError: `Segmentation fault` is detected by the operating system.
  [TimeInfo: *** Aborted at 1633948483 (unix time) try "date -d @1633948483" if you are using GNU date ***]
  [SignalInfo: *** SIGSEGV (@0x20) received by PID 55600 (TID 0x7fc772ffd700) from PID 32 ***]

尝试1：怀疑cuda版本过高导致，服务器cuda版本为11.1，查阅在GPU上如何选择PaddlePaddle版本？ (baidu.com)

决定先尝试将cuda版本降到10.2

wget https://developer.download.nvidia.com/compute/cuda/10.2/Prod/local_installers/cuda_10.2.89_440.33.01_linux.run
sudo sh cuda_10.2.89_440.33.01_linux.run

解决方案：

尝试1：成功