PyTorch C++ API missing headers?


I am a newer user of pytorch, and am currently constructing a CNN. I am using a Sequential object with a number of convolutional layers, and as such want to add some max pooling. However, my editor (CLion) tells me that the torch::nn::MaxPool1D class does not exist. I looked at the docs, and apparently it is supposed to be defined in a file called ....nn/modules/pooling.h, which does not appear in my copy of the sources. I have many others in the same directory, such as conv, dropout and more.
I am able to use torch::nn::Conv1d and torch::nn::Functional successfully, so what could be the reason for missing the pooling layers?

Kind regards

Are you using PyTorch 1.3? torch::nn::MaxPool1d should be in the 1.3 version.

The build-version file tells me I have version 1.3.0.dev20190820+cpu, which is a slightly older version I believe, but its still 1.3.0.

I just went through the trouble of upgrading to the most recent stable version which I got from the home-page (The one that I downloaded from the installation help page was nightly version which did not compile). I checked with the source code on github that it indeed contained the Pooling functions. Glad to confirm that it does indeed work now. Thanks for the help!