wangzixiang
/
lerobot
mirror of https://github.com/huggingface/lerobot.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
547 Commits 131 Branches 0 Tags 244 MiB
Commit Graph

132 Commits

Author SHA1 Message Date
Radek Osmulski 5d7a4e316b Minor change to improve test clarity 2024-05-30 07:36:58 +10:00
Radek Osmulski 621f69d98f Implement 2nd round of review changes 2024-05-30 07:36:58 +10:00
Radek Osmulski 566a8aa98e Implement review feedback 2024-05-30 07:36:58 +10:00
Radek Osmulski e7abcc2ffd add EpisodeAwareSampler 
EpisodeAwareSampler allows the dropping of first n or last n frames
from each episode.

Additionally, you can select which episodes to run on by passig in
their ids (complements functionality of the hf dataset slice syntax
to give extra flexibility to the user).

* modify train.py and diffusion.yaml to leverage new functionality
  and train on pusht with last 7 steps per episode dropped
* add unit tests
 2024-05-30 07:36:58 +10:00
Alexander Soare e3b9f1c19b
Add resume training (#205) 
Co-authored-by: Remi <re.cadene@gmail.com>
 2024-05-28 12:04:23 +01:00
Alexander Soare e67da1d7a6
Add tutorials for using the training script and (#196) 
Co-authored-by: Remi <re.cadene@gmail.com>
 2024-05-21 16:47:49 +01:00
Remi c4da689171
Hot fix to compute validation loss example test (#200) 
Co-authored-by: Alexander Soare <alexander.soare159@gmail.com>
 2024-05-20 18:30:11 +02:00
Radek Osmulski 9b62c25f6c
Adds split_by_episodes to LeRobotDataset (#158) 2024-05-20 14:04:04 +02:00
Remi 01eae09ba6
Fix aloha real-world datasets (#175) 2024-05-20 13:48:09 +02:00
Alexander Soare 4d7d41cdee
Fix act action queue (#185) 2024-05-16 15:43:25 +01:00
Akshay Kashyap c9069df9f1
Port SpatialSoftmax and remove Robomimic dependency (#182) 
Co-authored-by: Alexander Soare <alexander.soare159@gmail.com>
 2024-05-16 15:34:10 +01:00
Alexander Soare 68c1b13406
Make policies compatible with other/multiple image keys (#149) 2024-05-16 13:51:53 +01:00
Simon Alibert f52f4f2cd2
Add copyrights (#157) 2024-05-15 12:13:09 +02:00
Simon Alibert 29a196c5dd
Fix #173 - Require gym-pusht to be installed for test_examples_3_and_2 (#174) 2024-05-12 08:08:59 +02:00
Alexander Soare 1249aee3ac
Enable logging all the information returned by the `forward` methods of policies (#151) 2024-05-10 07:45:32 +01:00
Alexander Soare b187942db4
Add context manager for seeding (#164) 2024-05-09 17:58:39 +01:00
Alexander Soare 473345fdf6
Fix stats override in ACT config (#161) 2024-05-09 15:16:47 +01:00
Alexander Soare e89521dfa0
Enable tests for TD-MPC (#160) 2024-05-09 13:42:12 +01:00
Alexander Soare a8e245fb31
Remove loss masking from diffusion policy (#135) 2024-05-06 07:27:01 +01:00
Simon Alibert f5e76393eb
Release cleanup (#132) 
Co-authored-by: Kashif Rasul <kashif.rasul@gmail.com>
Co-authored-by: Alexander Soare <alexander.soare159@gmail.com>
Co-authored-by: Adil Zouitine <adilzouitinegm@gmail.com>
Co-authored-by: Cadene <re.cadene@gmail.com>
 2024-05-06 03:03:14 +02:00
Remi 6eaffbef1d
Add 25 real world datasets (static + mobile aloha) (#133) 2024-05-05 19:13:11 +02:00
Alexander Soare f6fe162432
Rename Diffusion Policy model repository (#136) 2024-05-05 12:06:58 +01:00
Alexander Soare f3bba0270d
Remove EMA model from Diffusion Policy (#134) 2024-05-05 11:26:12 +01:00
Simon Alibert c77633c38c
Add regression tests (#119) 
- Add `tests/scripts/save_policy_to_safetensor.py` to generate test artifacts
- Add `test_backward_compatibility to test generated outputs from the policies against artifacts
 2024-05-04 16:20:30 +02:00
Remi 19812ca470
Add dataset visualization with rerun.io (#131) 
Co-authored-by: Simon Alibert <75076266+aliberts@users.noreply.github.com>
 2024-05-04 16:07:14 +02:00
Alexander Soare bccee745c3
Refactor eval.py (#127) 2024-05-03 17:33:16 +01:00
Remi b2cda12f87
Add video decoding to LeRobotDataset (#92) 2024-05-03 00:50:19 +02:00
Alexander Soare d1855a202a
Refactor TD-MPC (#103) 
Co-authored-by: Cadene <re.cadene@gmail.com>
Co-authored-by: Simon Alibert <75076266+aliberts@users.noreply.github.com>
 2024-05-01 16:40:04 +01:00
Alexander Soare a4891095e4
Use PytorchModelHubMixin to save models as safetensors (#125) 
Co-authored-by: Remi <re.cadene@gmail.com>
 2024-05-01 16:17:18 +01:00
Alexander Soare 01d5490d44
Follow `transformers` single file naming conventions (#124) 2024-05-01 13:09:42 +01:00
Alexander Soare 9d60dce6f3
Tidy up yaml configs (#121) 2024-04-30 16:08:59 +01:00
Remi e4e739f4f8
Refactor push_dataset_to_hub (#118) 2024-04-30 14:25:41 +02:00
Quentin Gallouédec 508bd92d03
Remove `update` method from the policy (#99) 
Co-authored-by: Alexander Soare <alexander.soare159@gmail.com>
 2024-04-29 12:27:58 +02:00
Alexander Soare ccffa9e406
Fix policy defaults (#113) 2024-04-29 08:26:59 +01:00
Adil Zouitine 55dc9f7f51
Refactor the download and publication of the datasets and convert it into CLI script (#95) 
Co-authored-by: Remi <re.cadene@gmail.com>
 2024-04-29 00:08:17 +02:00
Adil Zouitine 81e490d46f
Add UMI-gripper dataset (#83) 
Co-authored-by: Remi <re.cadene@gmail.com>
 2024-04-28 18:41:07 +02:00
Simon Alibert bf2eebb090
Update readme & remove example 1 (#108) 
Co-authored-by: Remi <re.cadene@gmail.com>
- Update instructions for installing the library
- Remove deprecated example 1 (as we are now only using `LeRobotDataset` since #91)
 2024-04-27 09:48:02 +02:00
Simon Alibert fdf6a0c4e3
More CI cleanup, add style workflow (#107) 
- Changes on the `test.yml` workflow:
  - Using poetry instead of pip. Contrary to what I wrote in #75, it is possible to use poetry (and have the benefits of shorter install times) without the need for having two separate versions of `pyproject.toml` and `poetry.lock`.
  - Reduce the trigger scope to only run when files in these directories are modified:
    - `lerobot/`
    - `tests/`
    - `examples/`
    - `.github/`
- Add `style.yml` workflow for doing a `ruff check` pass on the code
- More cleanup (removed deprecated workflow)
 2024-04-27 09:37:56 +02:00
Simon Alibert b980c5dd9e
CI nightlies cpu/gpu & cleanup (#75) 2024-04-25 14:58:39 +02:00
Remi 659c69a1c0
Refactor datasets into LeRobotDataset (#91) 
Co-authored-by: Alexander Soare <alexander.soare159@gmail.com>
 2024-04-25 12:23:12 +02:00
Remi e760e4cd63
Move normalization to policy for act and diffusion (#90) 
Co-authored-by: Alexander Soare <alexander.soare159@gmail.com>
 2024-04-25 11:47:38 +02:00
Remi 1030ea0070
Loads episode_data_index and stats during dataset __init__ (#85) 
Co-authored-by: Simon Alibert <75076266+aliberts@users.noreply.github.com>
Co-authored-by: Alexander Soare <alexander.soare159@gmail.com>
 2024-04-23 14:13:25 +02:00
Simon Alibert 7ad1909641
Tests cleaning & simplification (#81) 2024-04-18 14:47:42 +02:00
Remi 0928afd37d
Improve dataset examples (#82) 
Co-authored-by: Alexander Soare <alexander.soare159@gmail.com>
 2024-04-18 11:43:16 +02:00
Alexander Soare bff4b673c9 Merge remote-tracking branch 'upstream/main' into unify_policy_api 2024-04-17 08:08:57 +01:00
Cadene 91badebdfc fix tests 2024-04-16 17:29:31 +00:00
Cadene 36d9e885ef Address comments 2024-04-16 17:20:54 +00:00
Cadene 4a3eac4743 fix unit tests, stats was missing, visualize_dataset was broken 2024-04-16 17:20:53 +00:00
Cadene 0980fff6cc HF datasets works 2024-04-16 17:19:40 +00:00
Cadene c6aca7fe44 For Pusht: use hf datasets to train, rename load_data_with_delta_timestamps -> load_previous_and_future_frames 2024-04-16 17:19:06 +00:00