DataLoader time is always 0
I run your example https://github.com/pytorch/kineto/blob/main/tb_plugin/examples/resnet50_profiler_api.py and I keep getting a 0ms spent on the DataLoder. I tried to add a slow collate_fn function to see if that changed but I still get 0.
Is this the expected behaviour?
Issue → PyTorch profiler not capturing Dataloader time and runtime. Always shows 0. Code used → I have used the code given in official PyTorch profiler documentation ( PyTorch documentation)
Hardware Used-> Nvidia AI100 gpu
PyTorch version-> 1.13.0+cu117
PyTorch tensorboard profiler version → 0.4.1

hi @aaronenyeshi , can you please help me here
I encounter the same issue since pytorch-1.12.1 (pytorch-1.13.0 and pytorch-2.0.0 are faulty too). Everything worked well with pytorch-1.11.0. Also, only one step is displayed in the overview tab even though I profiled N steps.
@guotuofeng I have same issue. Could you please help me ?