Add world_size dict getter method for simple integration with W&B #1735
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Thanks for this amazing distributed learning tool for users!
I have something to propose for integration with W&B.
What
This PR adds two getter methods to
megatron.core.parallel_state.py
:get_all_world_sizes()
: returns a string of world sizes for TP, DP, CP, PP, and EP.get_world_sizes_dict()
: returns the same information as a dictionary, e.g.{"tp": 4, "dp": 8, "cp": 1, "pp": 2, "ep": 1}
.Why
When integrating Megatron-LM with logging tools such as Weights & Biases, it is often useful to record the parallelism configuration (world sizes for each parallel group) alongside the training metrics.
Currently, there is a convenient
get_all_ranks()
method for ranks, but no equivalent for world sizes.Benefits
Thanks for your time for reviewing this PR.
Thanks.