ubuntu20.04，安装tensorRT, 执行example里面的./sample_char_rnn程序，测试时候报了如标题的一个错误，居然如下两行代码这样解决了，这两行命令好像是重新加载nvidia内核模块，有点玄学：

sudo rmmod nvidia_uvm
sudo modprobe nvidia_uvm

错误详情：

TensorRT-8.6.1.6/bin$ ./sample_char_rnn
&&&& RUNNING TensorRT.sample_char_rnn [TensorRT v8601] # ./sample_char_rnn
[11/26/2023-22:48:00] [I] Building and running a GPU inference engine for Char RNN model...
[11/26/2023-22:48:00] [I] Done reading weights from file...
[11/26/2023-22:48:00] [W] [TRT] Unable to determine GPU memory usage
[11/26/2023-22:48:00] [W] [TRT] Unable to determine GPU memory usage
[11/26/2023-22:48:00] [I] [TRT] [MemUsageChange] Init CUDA: CPU +0, GPU +0, now: CPU 21, GPU 0 (MiB)
[11/26/2023-22:48:00] [W] [TRT] CUDA initialization failure with error: 999. Please check your CUDA installation:  http://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html
&&&& FAILED TensorRT.sample_char_rnn [TensorRT v8601] # ./sample_char_rnn