torchtune icon indicating copy to clipboard operation
torchtune copied to clipboard

[Multi-Node] Test ``recipes/knowledge_distillation_distributed.py`` on a multi-node setup

Open joecummings opened this issue 10 months ago • 2 comments

Goal

Confirm that recipes/knowledge_distillation_distributed.py works on a multi-node setup

Validation

Run the recipe, confirm the logs are correct. Prefer TensorBoard or Weights&Biases charts, as well.

Artifacts

If everything works as expected, please submit a PR updating the chart. If you hit any errors or edge cases, please feel free to fix them in your PR; however, if you don't want to do that you can also update this Github Issue with the problems for someone else to work on.

joecummings avatar Apr 29 '25 15:04 joecummings

I will take this up

Vishnu-sai-teja avatar May 02 '25 03:05 Vishnu-sai-teja

I will take this up

Thanks! Tag me on the PR when you're done :)

joecummings avatar May 02 '25 14:05 joecummings