Fix overflow in nanosets with big datasets #182
Conversation
This file contains 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
NouamaneTazi
requested changes
May 27, 2024
| input_ids_tokens = np.frombuffer( | ||
| self.dataset_buffer, dtype=self.token_dtype, count=(self.sequence_length + 1), offset=int(offset) | ||
| self.dataset_buffer, dtype=self.token_dtype, count=(self.sequence_length + 1), offset=offset |
There was a problem hiding this comment.
Nice find! Can you add a small unittest for this in tests/nanoset/test_build_nanoset_dataloader.py? thanks!
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Outdated suggestions cannot be applied.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.
When a nanoset is particularly big (>4 GB), the calculation of
offset(the actual location within the memmap) can overflow. The issue is with the lineoffset = dataset_sample * self.sequence_length * (np.iinfo(self.token_dtype).bits / 8)Here,
dataset_sampleis a numpy"uint", and the calculation ofoffsetcan overflow since numpy's"uint"type is 32 bits. The solution is to promote everything to native Pythonintfirst, which has automatic overflow.