Repository navigation
Conversation
aliberts
force-pushed
the
user/aliberts/2024_11_25_compute_stats_v2
branch
from
February 9, 2025 13:26
8feeede to
0c55461
Compare
Cadene
changed the base branch from
user/aliberts/2024_11_25_compute_stats_v2
to
user/aliberts/2025_02_10_dataset_v2.1
February 10, 2025 15:51
Cadene
force-pushed
the
user/rcadene/2025_01_27_dataset_v2.1
branch
from
February 11, 2025 15:53
34f201c to
8e98c79
Compare
Cadene
marked this pull request as ready for review
February 11, 2025 16:14
Cadene
force-pushed
the
user/rcadene/2025_01_27_dataset_v2.1
branch
from
February 11, 2025 16:20
ef18a95 to
3281e66
Compare
aliberts
approved these changes
Feb 13, 2025
aliberts
left a comment
Contributor
There was a problem hiding this comment.
LGTM with some comments
Comment on lines
+222
to
+223
| task_index = self.get_task_index(task) | ||
| if task_index not in self.tasks: |
Contributor
There was a problem hiding this comment.
Equivalent but a bit faster and easier to read I think
Suggested change
| task_index = self.get_task_index(task) | |
| if task_index not in self.tasks: | |
| if task not in self.tasks.values(): | |
| task_index = self.get_task_index(task) |
Contributor
Author
There was a problem hiding this comment.
I found this logic quite confusing, so I refactor it.
| dataset.add_frame({"1d": torch.randn(1)}) | ||
|
|
||
|
|
||
| def test_add_frame(tmp_path): |
Contributor
There was a problem hiding this comment.
This test is quite long for what it's doing (~0.5s)
Did you check if this new add_frame is slower than before?
Contributor
Author
|
reverting commit: Because getting: |
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.

What this does
Motivations: We want to start supporting datasets with multiple tasks per episode. We also want to iterate on a simpler API for adding new frames to a dataset, where all features are added using
dataset.add_frame(frame: dict).For instance:
TODO
taskargument ofsave_episode.add_frameis now the only way to add the task (even if it's the same for every frame)add_frameexamples/port_datasets/pusht_zarr.pyaccordinglyHow it was tested
python examples/port_datasets/pusht_zarr.pyHow to checkout & try? (for the reviewer)