GPU_VERSION #1
Description
Thank you for releasing this very useful, biology-informed deep-learning framework! I’m planning to train ExpiMap on a large single-cell dataset and would like to leverage my NVIDIA GPU to speed things up.torch.cuda.is_available() returns True and my other PyTorch scripts run on the GPU without issues.
However, when I instantiate an sca.models.EXPIMAP object and start train(), the model stays on the CPU even after explicitly calling
Does ExpiMap currently support GPU training out of the box, or is it CPU-only at the moment?