Created by: thunder95

   1）PaddlePaddle版本：1.5    2）CPU：i7    3）GPU：RTX2070    4）系统环境：ubuntu 16.04

使用官网提供的paddle trt的c++预测代码，能够顺利运行。下载地址

使用paddle c++预测自己训练的stnet模型，若不用trt也能顺利运行，但是一旦开启trt加速之后，就会报错。可能是输入的shape超过4维的原因，请问有什么解决办法？

报错内容： I1120 14:09:05.386106 12849 tensorrt_subgraph_pass.cc:270] Prepare TRT engine (Optimize model structure, Select OP kernel etc). This process may cost a lot of time. terminate called after throwing an instance of 'paddle::platform::EnforceNotMet' what(): Enforce failed. Expected shape.size() <= 4UL, but received shape.size():5 > 4UL:4. TensorRT' tensor input requires at most 4 dimensions at [/Paddle/paddle/fluid/inference/tensorrt/convert/op_converter.h:54]

输入的维度是这样： int batch_size=1; int segnum = 7; int channels = 15; //3*5 int height = 224; int width=224; input_t->Reshape({batch_size, segnum, channels, height, width});