Closed ningelschlingel closed 2 years ago
In order to running on some embedded devices, this implementation only use some normal layers or ops, and not bind to any special version libs, so I think it can be easily run at higher verison of PyTorch,you can have a try.
I got it to work now, just for reference, how long did the training take with coda enabled?
A few days ago, pytorch reportedly added support for apple metal: https://pytorch.org/blog/introducing-accelerated-pytorch-training-on-mac/
I am not sure yet if my setup is correct, because training takes very long. I updated all python requirements to the latest version and disabled CUDA with setting the GPU_ID in the config.py file to a false value. Some used functions seem to be deprecated now, but it runs
Thanks you for replying!
Support for GPU-accelerated training on Apple Silicon is apparently only available as of PyTorch version 1.12
I also have other problems with running it, and before I spend hours on trying to fix it for my setup, did anybody already do it?
Sorry for already filing an issue, if there was a Q/A section I would have used that before.
thanks in advance