there is no way to do this in pytorch. However, PyTorch doesn’t pre-occupy the GPU’s entire memory, so if your computation only uses 50% of GPU, only that much is locked by PyTorch
2 Likes
there is no way to do this in pytorch. However, PyTorch doesn’t pre-occupy the GPU’s entire memory, so if your computation only uses 50% of GPU, only that much is locked by PyTorch