hyperdimensional-computing / torchhd

Torchhd is a Python library for Hyperdimensional Computing and Vector Symbolic Architectures
https://torchhd.readthedocs.io
MIT License
238 stars 24 forks source link

Level embedding incompatible with Pytorch DDP #140

Closed rishikanthc closed 1 year ago

rishikanthc commented 1 year ago

So I have a model where the outputs of level embedding are being fed to a deep learning model. I'm trying to train the model on multiple GPUs using DDP but face the following error:

RuntimeError: Modules with uninitialized parameters can't be used withDistributedDataParallel. Run a dummy forward pass to correctly initialize the modules

The error occurs only when using Distributed Data Parallel (DDP). Works fine when using Data Parallel (DP).

mikeheddes commented 1 year ago

Hi, thanks for opening this issue. Could you provide a minimal example to recreate the error?

mikeheddes commented 1 year ago

I am closing this issue due to inactivity. Feel free to re-open if the issue has not been resolved.