Replies: 1 comment
-
不建议使用onnxruntime-gpu版,除非输入图像的尺寸固定。 |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
我想在GPU使用onnxruntime, 然后使用pip安装了onnxruntime-gpu==1.16.1,修改了TableStructureRec/wired_table_rec/utils.py的代码。
修改了utils.py下的OrtInferSession类,如下:
providers = ['CUDAExecutionProvider'] self.session = InferenceSession(str(model_path), sess_options=self.sess_opt, providers=providers)
之后运行demo后,发现推理速度变得更慢,且推理时显存会涨到18G。在CPU上运行正常。请问这个问题该怎么解决?
运行机器配置如下:
显卡:V100 32G
系统:ubuntu18.04
cpu:Intel(R) Xeon(R) Platinum 8163 CPU @ 2.50GHz(12核)
cuda:11.7
onnxruntime-gpu:1.16.1
Beta Was this translation helpful? Give feedback.
All reactions