fix(iterable): ensure MappedExamplesIterable supports state_dict for resume #7656
Conversation
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
…resume Fixes huggingface#7630 ### Problem When calling `.map()` on an `IterableDataset`, resuming from a checkpoint skips a large number of samples. This is because `MappedExamplesIterable` did not implement `state_dict()` or `load_state_dict()`, so checkpointing was not properly delegated to the underlying iterable. ### What This PR Does This patch adds: ```python def state_dict(self): return self.ex_iterable.state_dict() def load_state_dict(self, state): self.ex_iterable.load_state_dict(state) to MappedExamplesIterable, so the wrapped base iterable's state can be saved and restored as expected. Result Using .map() no longer causes sample skipping after checkpoint resume. Tested manually using the reproducer script in huggingface#7630. The resumed iterator now yields the correct samples. Let me know if a dedicated test case is required — happy to add one!
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Outdated suggestions cannot be applied.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.
Fixes #7630
Problem
When calling
.map()on anIterableDataset, resuming from a checkpoint skips a large number of samples. This is becauseMappedExamplesIterabledid not implementstate_dict()orload_state_dict(), so checkpointing was not properly delegated to the underlying iterable.What This PR Does
This patch adds:
to MappedExamplesIterable, so the wrapped base iterable's state can be saved and restored as expected.
Result
Using .map() no longer causes sample skipping after checkpoint resume.
Let me know if a dedicated test case is required — happy to add one!