Add option for custom filtering based on json#621
Merged
kpertsch merged 12 commits intoPhysical-Intelligence:mainfrom Aug 31, 2025
Merged
Add option for custom filtering based on json#621kpertsch merged 12 commits intoPhysical-Intelligence:mainfrom
kpertsch merged 12 commits intoPhysical-Intelligence:mainfrom
Conversation
verityw
requested review from
Michael-Equi,
jimmyt857,
kvablack and
uzhilinsky
as code owners
kpertsch
approved these changes
Aug 30, 2025
kpertsch
reviewed
Aug 30, 2025
ChenZhenY
pushed a commit
to ChenZhenY/openpi
that referenced
this pull request
Sep 23, 2025
…l-Intelligence#621) The current training code only uses the default filter. This PR adds the option to filter based off of JSONs that map episodes to ranges of timesteps to keep. - Changed `config.py` to also have the option to specify where that `json` filter file is. - Changed `data_loader.py` to make use of above. - Changed `DroidRldsDataset` to load that `json`, then filter out all frames not contained within that file. - Added `examples/droid/get_droid_keep_ranges.py`, which uses a more aggressive filter (created by Karl) to generate such a json. - Updated `README_train.md` to reflect these changes. Added reference to DROID hf repo, which contains a copy of that json run on `droid/1.0.1`.
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The current training code only uses the default filter. This PR adds the option to filter based off of JSONs that map episodes to ranges of timesteps to keep.
config.pyto also have the option to specify where thatjsonfilter file is.data_loader.pyto make use of above.DroidRldsDatasetto load thatjson, then filter out all frames not contained within that file.examples/droid/get_droid_keep_ranges.py, which uses a more aggressive filter (created by Karl) to generate such a json.README_train.mdto reflect these changes. Added reference to DROID hf repo, which contains a copy of that json run ondroid/1.0.1.