Kazuhiro Yamasaki
Kazuhiro Yamasaki
**Describe the bug** Dear ESPnet team, This issue is a proposal to improve the multi-GPU training performance for **v1** ASR. I'm filing this because I didn't find any other issues...
I found an issue about model training loss in [DIBR tutorial](https://github.com/NVIDIAGameWorks/kaolin/blob/master/examples/tutorial/dibr_tutorial.ipynb). Could you check more details? ## Problem In contrast to an example log message on tutorial notebook, model training...
Plan for new markers representing data I/O and communication between GPUs.
On jupyter notebook, it's better for every developers/researchers to get profiling result for each cell. Try to support jupyter notebook.