Added Intermediate Tutorial on Distributed Training with CIFAR10

pytorch-ignite / examples

Examples, tutorials, and how-to guides

BSD 3-Clause "New" or "Revised" License

11 stars 12 forks source link

Added Intermediate Tutorial on Distributed Training with CIFAR10 #26

Closed Priyansi closed 3 years ago

Priyansi commented 3 years ago

Converted CIFAR10 Distributed example into an intermediate tutorial with explanation for every function and outputs for different configurations.

To do:

[x] Check if the tutorial is running on single node, multiple GPUs and multiple node, multiple GPUs.

Priyansi commented 3 years ago

@sdesrozis I have made the suggested changes!

Priyansi commented 3 years ago

@vfdev-5 I have made the suggested changes!

vfdev-5 commented 3 years ago

In the end of the tutorial let's put links to ClearML experiments to explore the runs.

Priyansi commented 3 years ago

@vfdev-5 can you please review this again?

sdesrozis commented 3 years ago

@Priyansi I left some comments.

Priyansi commented 3 years ago

Thank you for the suggestions @sdesrozis and @vfdev-5 . I left a few comments as well.

Priyansi commented 3 years ago

@vfdev-5 I added the note for TPUs and included the output for both TPUs and Jupyter Notebooks.

vfdev-5 commented 3 years ago

@Priyansi thanks for all the updates ! There are few other things to update in above comments. In the end I think we have to provide a link on complete code:

cifar10-distributed.py near the notebook in the folder
Links to ClearML experiments (https://github.com/pytorch-ignite/examples/pull/26#issuecomment-913455194). Let's do that once PR is merged and we'll update the notebook