How we can potentially support CUDA accelerator in the future #1
LazaroHurtado
started this conversation in
Ideas
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Below are some introductory resources for getting familiar with CUDA.
In the future, it would be great to support CUDA devices since training on CPU is extremely inefficient. Accelerators would come into play during the forward and backward pass. Thus, refactoring of the ops module would most likely be required for backward pass along side adding support for automatic kernel accelerator selection when applying the forward pass on a
Tensor<_>
.This is a developing discussion.
Beta Was this translation helpful? Give feedback.
All reactions