Initial steps toward a torch.utils.data.IterableDataset (PR 3 of N) #6
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
ryan-williams
force-pushed
the
bkmartinrj/delete-old-files
branch
2 times, most recently
from
1cb7db7
to
54ae565
Compare
ryan-williams
force-pushed
the
bkmartinjr/initial-non-shuffling-code
branch
from
4a546ca
to
c0f32a1
Compare
ryan-williams
added a commit
that referenced
this pull request
Sep 25, 2024
(to enable testing on #6)
* add GHA workflows; remove temp exclusion from pre-commit config * GHA tweak, attempt to trigger PR runs --------- Co-authored-by: Ryan Williams <ryan.williams@tiledb.com>
ryan-williams
force-pushed
the
bkmartinjr/initial-non-shuffling-code
branch
from
c0f32a1
to
d9c0ef4
Compare
Merged
There was a problem hiding this comment.
Most of the way through, couple q's here, and suggested changes in #15!
src/tiledbsoma_ml/pytorch.py
Outdated
| assert self._var_joinids is not None | ||
| world_size, _ = _get_distributed_world_rank() | ||
| n_workers, _ = _get_worker_world_rank() | ||
| partition_len = len(self._obs_joinids) // world_size // n_workers |
There was a problem hiding this comment.
Will some partitions be one greater than others, if this division has remainders? Does it matter?
There was a problem hiding this comment.
Yes, they can vary across data loader workers. And it doesn't really matter. What matters the most is having the same number of samples distributed to each GPU. Beyond that, things are data and config dependent.
ryan-williams
added a commit
that referenced
this pull request
Oct 1, 2024
ryan-williams
force-pushed
the
bkmartinjr/initial-non-shuffling-code
branch
from
cd51799
to
579727a
Compare
…rse_output__batched`
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.
This PR adds a simple
torch.utils.data.IterableDatasetandtorchdata.datapipes.iter.IterDataPipeimplementation. An initial set of tests is included.Notes to reviewer:
torch.distributedvariants of world and rank.dataloaderconvenience API - that will come in a later PR