search

serizba / cppflow

Run TensorFlow models in C++ without installation and without Bazel
https://serizba.github.io/cppflow/
MIT License
787 stars 179 forks source link

How do you set the device? #174

Closed ghost closed 2 years ago

ghost commented 2 years ago

When it creates a tensor it always puts it on CPU. Despite I can copy it to GPU, it still has a foreboding warning that changes to that TF_Tensor do not reflect the actual device. So, my tensor claims I copied it to GPU but it is still peforming worse than CPU. Even Eigen is way faster than this thing.

ghost commented 2 years ago

I turn on log placement and everything is on the GPU. I can conclude something wrong with my codes, it is not a cppflow problem.