Hello! It seems that the currently released code can run training and inference without relying on Pytorch Geometric. Also looking at all the imports in the code I did not observe anything from PyG. Is there a faster version of the code that's not (or yet to be) released that relies on PyG? Thanks!
Hello! It seems that the currently released code can run training and inference without relying on Pytorch Geometric. Also looking at all the imports in the code I did not observe anything from PyG. Is there a faster version of the code that's not (or yet to be) released that relies on PyG? Thanks!