Flaviu
Flaviu
Workaround is to load native binaries yourself and run the empty_cache method. Like it has been done here for example: https://github.com/K024/llm-sharp
```c# var gpu0 = new Device(DeviceType.CUDA, 0); var gpu1 = new Device(DeviceType.CUDA, 1); var myTensor0 = torch.ones(3, 3).to(gpu0); var myTensor1 = torch.ones(3, 3).to(gpu1); ... ``` I think this should work.
I hope the new release will support cuda 12.8 because we are planing on switching to RTX50 cards soon.
What is the status on this one? I waiting for this to be merged.