When running test.py, I am experiencing significantly slow performance. I have verified that both the data and the model are on cuda:0. However, I observed that the GPU memory on cuda:0 is not being utilized as expected. Could this be a configuration issue, or is there something else I might be missing?
When running test.py, I am experiencing significantly slow performance. I have verified that both the data and the model are on cuda:0. However, I observed that the GPU memory on cuda:0 is not being utilized as expected. Could this be a configuration issue, or is there something else I might be missing?