Repository navigation
Fix DSV4 DSpark sample-from-anchor initialization - #36419
Merged
hnyls2002 merged 2 commits intoAug 26, 2026
Merged
Conversation
Root cause: the shared DSpark runner now requires every draft model to expose sample_from_anchor, but the standalone DeepSeek-V4 DSpark implementation did not inherit the generic DSpark mixin that initializes it. Startup therefore failed with AttributeError.\n\nFix: initialize sample_from_anchor directly from the DeepSeek-V4 draft checkpoint config using the shared DSpark config helper.\n\nValidation: Python syntax compilation, diff checks, and the full pre-commit suite passed.
Contributor
Author
|
@Fridge003 hi, could you have a look? |
Contributor
Author
|
/tag-and-rerun-ci |
Collaborator
|
/rerun-test test_deepseek_v4_shared_expert_fusion.py test_runai_model_streamer_loader.py test_dspark_draft_path_default.py |
Contributor
|
Results for 🚀 |
5 tasks done
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Motivation
The shared DSpark runner reads sample_from_anchor from every draft model. The generic DSpark mixin initializes it, but DeepseekV4ForCausalLMDSpark is a standalone implementation and omitted the attribute, causing startup to fail with AttributeError after #36186.
Validation
CI States
Latest PR Test (Base): ⏳ Run #32932193005
Latest PR Test (Extra): ❌ Run #32932192777
Latest PR Test (AMD ROCm 7.2): ⏳ Run #32932192944