fix: prevent duplicate expert IDs on same GPU (#22)#27
Open
yurekami wants to merge 2 commits into
Open
Conversation
Add proper Python packaging support to enable pip installation: - Add pyproject.toml with modern build system configuration - Add module docstring describing the package - Add __version__ = "1.0.0" for version tracking - Expand __all__ to export additional utility functions This enables users to install the package via: pip install git+https://github.com/deepseek-ai/EPLB.git Or for local development: pip install -e . 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Modified balanced_packing() to accept an optional item_ids parameter that prevents items with the same ID from being placed in the same pack. When packing physical experts to GPUs in rebalance_experts_hierarchical(), pass phy2mlog as item_ids to ensure no two physical experts that map to the same logical expert end up on the same GPU. This fixes the issue where hot experts could be duplicated on the same rank, e.g., "dup expert layer 0 rank 16, [200, 226, 236, 236, 12, ...]" 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Fix the issue where hot experts could be duplicated on the same rank, e.g., "dup expert layer 0 rank 16, [200, 226, 236, 236, 12, ...]"
Changes
Modified
balanced_packing()to accept an optionalitem_idsparameter that prevents items with the same ID from being placed in the same pack.When packing physical experts to GPUs in
rebalance_experts_hierarchical(), passphy2mlogas item_ids to ensure no two physical experts that map to the same logical expert end up on the same GPU.Test plan
🤖 Generated with Claude Code