Skip to content

Sync with Lerobot upstream for 101 use #9

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 202 commits into from
Jul 23, 2025
Merged

Sync with Lerobot upstream for 101 use #9

merged 202 commits into from
Jul 23, 2025

Conversation

@AmanuelErgogo
Copy link

@AmanuelErgogo AmanuelErgogo commented Jul 21, 2025 •
edited
Loading

1. Relevant issue

closes https://github.com/TheRARELab/robot-learning-project/issues/173

2. Describe/summarize the changes

Syncing our main with Lerobot upstream

3. Tell why the changes in this PR work or high-quality

It resembles the existing documentation but specifies the download from Hugging face and exporting format

PR creator's Responsibility Checklist

  • I have done the following
    1. followed the branch-based coding workflow, including single-purpose branch and single-purpose PR
    2. routed lab GitHub notifications to my usf.edu inbox to receive email notifications and will
    3. watched all repos under my GitHub team and set it to watch new repos automatically
  • I have summarized the changes in the title, which is not the space-separated branch name.
  • I have filled out all three required sections.
  • I have included only the files and lines relevant to this change.
  • I have reviewed each change at the word level to ensure high-quality work.
  • If applicable, I have revised or added documentation (readme, code comments) so others can understand & reproduce the changes.
    • If docs are added, I have described them in the "Summarize the changes" section above.
  • I will assign reviewer(s) under Reviewers in the PR sidebar, not Assignees, so they get email notifications (doc).

Reviewer's Responsibility Checklist

  • I understand how to review a PR (more comprehensive guide).
  • This PR closes a task. If not specified, PR creator explains it clearly.
  • The title & change description reflect what the changes intend to do, and I understand it.
  • I look at every line of non-generated code & understand what the code or other changes intend to do, including the naming of everything, comments, consistent style, and simple design.
  • I can reproduce the changes with existing or changed documentation: Setting up dev environment, running programs, and how to use.
  • Finally, I leave comments on encouragement and appreciation for good practices by the PR creator.

aliberts and others added 30 commits January 30, 2025 22:06
@aliberts
Comply with torchvision 0.21 custom transforms (huggingface#665)
1ee1acf
@aliberts @Cadene @tc-huang
Simplify configs (huggingface#550)
3c0a209 
Co-authored-by: Remi <[email protected]>
Co-authored-by: HUANG TZU-CHUN <[email protected]>
@pkooij
fix setting motor id with new dataclass config (huggingface#668)
ae9605f
@aliberts
Fix Docker cpu/gpu builds (huggingface#667)
6674e36
@aliberts
Fix nightly tests docker images (huggingface#675)
43e079f
@pkooij
User/pepijn/2025 01 31 improved tutorial so100 (huggingface#666)
dd97452
@Cadene @aliberts @molbap
Add Pi0 (huggingface#681)
638d411 
Co-authored-by: Simon Alibert <[email protected]>
Co-authored-by: Simon Alibert <[email protected]>
Co-authored-by: Pablo <[email protected]>
@aliberts
Remove dataset tests artifacts (huggingface#701)
8548a87
@aliberts
Update CI trigger rules (huggingface#712)
334deb9
@aliberts @Cadene
Remove offline training, refactor train.py and logging/checkpointing (
90e099b 
huggingface#670)

Co-authored-by: Remi <[email protected]>
@aliberts
Fixes following huggingface#670 (huggingface#719)
e710959
@aliberts
Prevent resuming from hub (huggingface#726)
1e49cc4
@IliaLarchenko
Fixed eval.py on MPS (huggingface#702)
c574eb4
@aliberts
Fix wandb logging (huggingface#730)
b86a2c0
@aliberts
Add zizmor pre-commit (huggingface#732)
2cb0bf5
@aliberts
Update pre-commits (huggingface#733)
c4c2ce0
@aliberts
Conform pyproject to PEP 621 (huggingface#621)
ddeade0
@aliberts @Cadene
Remove poetry.lock (huggingface#737)
fe483b1 
Co-authored-by: Remi <[email protected]>
@pkooij
Add mobile so100 (huggingface#724)
aca464c
@aliberts @Cadene
LeRobotDataset v2.1 (huggingface#711)
3354d91 
Co-authored-by: Remi <[email protected]>
Co-authored-by: Remi Cadene <[email protected]>
@Haskely @aliberts
feat: Add root directory option for dataset configuration (huggingfac…
eecf32e 
…e#765)

Co-authored-by: Simon Alibert <[email protected]>
@yongjincho
Fix the URL to setup hardware Aloha Stationary in the example document (
e64fad2 
huggingface#766)
@YoussefBayouli
Optimizing Dockerfile (huggingface#751)
75d5fa4
@raulgarreta
fixes in 7_get_started_with_real_robot.md (huggingface#677)
65db5af
@J4nn1K
[QOL] Enable teleoperation during environment reset (huggingface#725)
8699a28
@aliberts
Add typos checks (huggingface#770)
a1809ad
@aliberts
Add pr style bot (huggingface#772)
da265ca
@aliberts
Fix nightly (huggingface#775)
659ec44
@mishig25
[vizualisation] Add pagination for many episodes (huggingface#776)
8d60ac3
@aliberts
Fix pr_style_bot (huggingface#781)
68b369e
ben-z and others added 24 commits July 4, 2025 14:56
@ben-z
Fix indentation (huggingface#1436)
aec1b29
@fracapuano
Fixes @torch.no_grad() usage (huggingface#1455)
a5e0aae 
* fix: decorator calls with parentheses

* fix no grad for normalize too

Signed-off-by: Francesco Capuano <[email protected]>

---------

Signed-off-by: Francesco Capuano <[email protected]>
@aliberts @pkooij
@imstevenpmwork @pre-commit-ci @nepyope
Add Hope Jr (huggingface#935)
039de25 
* Fix imports

* Add feetech write tests

* Nit

* Add autoclosing fixture

* Assert ping stub called

* Add CalibrationMode

* Add Motor in dxl robots

* Simplify split_int_bytes

* Rename read/write -> sync_read/write, refactor, add write

* Rename tests

* Refactor dxl tests by functionality

* Add dxl write test

* Refactor _is_comm_success

* Refactor feetech tests by functionality

* Add feetech write test

* Simplify _is_comm_success & _is_error

* Move mock_serial patch to dedicated file

* Remove test skips & fix docstrings

* Nit

* Add dxl operating modes

* Add is_connected in robots and teleops

* Update Koch

* Add feetech operating modes

* Caps dxl OperatingMode

* Update ensure_safe_goal_position

* Update so100

* Privatize methods & renames

* Fix dict

* Add _configure_motors & move ping methods

* Return models (str) with pings

* Implement feetech broadcast ping

* Add raw_values option

* Rename idx -> id_

* Improve errors

* Fix feetech ping tests

* Ensure motors exist at connection time

* Update tests

* Add test_motors_bus

* Move DriveMode & TorqueMode

* Update Koch imports

* Update so100 imports

* Fix visualize_motors_bus

* Fix imports

* Add calibration

* Rename idx -> id_

* Rename idx -> id_

* (WIP) _async_read

* Add new calibration method for robot refactor (huggingface#896)

Co-authored-by: Simon Alibert <[email protected]>

* Remove deprecated scripts

* Rename CalibrationMode -> MotorNormMode

* Fix calibration functions

* Remove todo

* Add scan_port utility

* Add calibration utilities

* Move encoding functions to encoding_utils

* Add test_encoding_utils

* Rename test

* Add more calibration utilities

* Format baudrate tables

* Implement SO-100 leader calibration

* Implement SO-100 follower calibration

* Implement Koch calibration

* Add test_scan_port (TODO)

* Fix calibration

* Hack feetech firmware bug

* Update tests

* Update Koch & SO-100

* Improve format

* Rename SO-100 classes

* Rename Koch classes

* Add calibration tests

* Remove old calibration tests

* Revert feetech hack and monkeypatch instead

* Simplify motors mocks

* Add is_calibrated test

* Update viperx & widowx

* Rename viperx & widowx

* Remove old calibration

* feat(teleop): thread-safe keyboard teleop implementation (huggingface#869)

Co-authored-by: Simon Alibert <[email protected]>

* Add support for feetech scs series + various fixes

* Update dynamixel with motors bus & tables changes

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* (WIP) Add Hope Jr

* Rename arm -> hand

* (WIP) Add homonculus arm & glove

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Add Feetech protocol version

* Implement read

* Use constants from sdks

* (nit) move write

* Fix broadcast ping type hint

* Add protocol 1 broadcast ping

* Refactor & add _serialize_data

* Add feetech sm8512bl

* Make feetech broadcast ping faster in protocol 1

* Cleanup

* Add support for feetech protocol 1 to _split_into_byte_chunks

* Fix unormalize

* Remove test_motors_bus fixtures

* Add more segmented tests (base motor bus & feetech), add feetech protocol 1 support

* Add more segmented tests (dynamixel)

* Refactor tests

* Add handshake, fix feetech _read_firmware_version

* Fix tests

* Motors config & disconnect fixes

* Add torque_disabled context

* Update branch & fix pre-commit errors

* Fix hand & glove readings

* Update feetech tables

* Move read/write_calibration implementations

* Add setup_motor

* Fix calibration msg display

* Fix setup_motor & add it to robots

* Fix _find_single_motor

* Remove deprecated configure_motor

* Remove deprecated dynamixel_calibration

* Remove names

* Remove deprecated import

* refactor/lekiwi robot (huggingface#863)

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Simon Alibert <[email protected]>
Co-authored-by: Simon Alibert <[email protected]>

* fix(teleoperators): use property is_connected (huggingface#1075)

* Remove deprecated manipulator

* Update robot features & naming

* Update teleop features & naming

* Add make_teleoperator_from_config

* Rename find_port

* Fix config parsing

* Remove app script

* Add setup_motors

* Add teleoperate

* Add record

* Add replay

* Fix test_datasets

* Add mock robot & teleop

* Add new test_control_robot

* Add test_record_and_resume

* Remove deprecated scripts & tests

* Add calibrate

* Add docstrings

* Fix tests (no-extras install)

* Add SO101

* Remove pynput from optional deps

* Rename example 7

* Remove unecessary id

* Add MotorsBus docstrings

* Rename arm -> bus

* Remove Moss arm

* Fix setup_motors & calibrate configs

* Fix test_calibrate

* Add copyrights

* Update hand & arm

* Update homonculus hand & arm

* Fix dxl _find_single_motor

* Update glove

* Add setup_motors for lekiwi

* Fix glove calibration

* Complete docstring

* Add check for same min and max during calibration

* Move MockMotorsBus

* Add so100_follower tests

* (WIP) add calibration gui

* Fix test

* Add setup_motors

* Update calibration gui

* Remove old .cache folder

* Replace deprecated abc.abstractproperty

* Fix feetech protocol 1 configure

* Cleanup gui & add copyrights

* Anatomically precise joint names

* (WIP) Add glove to hand joints translation

* Move make_robot_config

* Add drive_mode & norm_mode in glove calibration

* Fix joints translation

* Fix normalization drive_mode

* nit

* Fix glove to hand conversion

* Adapt feetech calibration

* Remove pygame prompt

* Implement arm calibration (hacks)

* Better MotorsBus error messages

* Update feetech read_calibration

* Fix feetech test_is_calibrated

* Cleanup glove

* (WIP) Update arm

* Add changes from huggingface#1117

* refactor(cameras): cameras implementations + tests improvements (huggingface#1108)

Co-authored-by: Simon Alibert <[email protected]>
Co-authored-by: Simon Alibert <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

* Fix arm joints order

* Add timeout/event logic

* Fix arm & glove

* Fix predict_action from record

* fix(cameras): update docstring + handle sn when starts with 0 + update timeouts to more reasonable value (huggingface#1154)

* fix(scripts): parser instead of draccus in record + add __get_path_fields__() to RecordConfig (huggingface#1155)

* Left/Right sides + other fixes

* Arm fixes and add config

* More hacks

* Add control scripts

* Fix merge errors

* push changes to calibration, teleop and docs

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Move readme to docs

* update readme

Signed-off-by: Martino Russi <[email protected]>

* Add files via upload

Signed-off-by: Martino Russi <[email protected]>

* Update image sources

* Symlink doc

* Compress image

* Move image

* Update docs link

* fix docs

* simplify teleop scripts

* fix variable names

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Address code review

* add EMA to glove

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* integrate teleoperation for hand

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* update docs

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* import hopejr/homunculus in teleoperate

* update docs for teleoperate, record, replay, train and inference

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* chore(hopejr): address comments

* chore(hopejr): address coments 2

* chore(docs): update teleoperation instructions for the hand/glove

* fix(hopejr): calibration int + update docs

---------

Signed-off-by: Martino Russi <[email protected]>
Signed-off-by: Simon Alibert <[email protected]>
Co-authored-by: Pepijn <[email protected]>
Co-authored-by: Steven Palma <[email protected]>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: nepyope <[email protected]>
Co-authored-by: Martino Russi <[email protected]>
Co-authored-by: Steven Palma <[email protected]>
@CarolinePascal
fix(logging): Fixing logging levels (huggingface#1466)
cf86b93 
* fix(logging): Fixing logging levels, adding custom logging levels for console and file logging

* clean(typing): Adding typing in logging formatter, use proper getter for logging message
@AdilZouitine @s1lent4gnt @jpizarrom
fix(hil-serl): discrete critic send through network (huggingface#1468)
ce2b972 
Co-authored-by: Khalil Meftah <[email protected]>
Co-authored-by: jpizarrom <[email protected]>
@fracapuano @imstevenpmwork @michel-aractingi
Add Async Inference (huggingface#1196)
30c1610 
Co-authored-by: Steven Palma <[email protected]>
Co-authored-by: Michel Aractingi <[email protected]>
@fracapuano @michel-aractingi
Update async docs with blogpost (huggingface#1479)
abe51ee 
Co-authored-by: Michel Aractingi <[email protected]>
@fracapuano @michel-aractingi
fix(docs): Record-Upload failed? Don't panic! (huggingface#1478)
d2645cb 
* fix: add instruction to manually upload dataset

Signed-off-by: Francesco Capuano <[email protected]>

* fix: repo type is explicited

---------

Signed-off-by: Francesco Capuano <[email protected]>
Co-authored-by: Michel Aractingi <[email protected]>
@ben-z
Remove random noise injected by policy server (huggingface#1496)
519b761
@AdilZouitine @michel-aractingi
fix(mps): gradient exploding and nan loss issues with ACT (huggingfac…
91b110d 
…e#1490)

Co-authored-by: Michel Aractingi <[email protected]>
@aliberts
Fix tests (huggingface#1510)
724874e
@todateman @pre-commit-ci
fix(record): Improve OpenCV backend handling for Windows systems (hug…
1b878c9 
…gingface#1495)

* fix(record): Improve OpenCV backend handling for Windows systems

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Resolved ruff's E402 error (import statements not at the beginning of the file):
- Moved all import statements to the beginning of the file
- Defined _fix_opencv_backend() as a function
- Adjusted the timing of the fix call
- Code structure conforming to ruff

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix(record): Correct OpenCV backend for Windows systems

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix(opencv): Set OpenCV environment variable for Windows systems

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix(opencv): Refactor MSMF hardware transform environment variable setting for Windows

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
@pre-commit-ci @aliberts
[pre-commit.ci] pre-commit autoupdate (huggingface#1327)
c4c0105 
* [pre-commit.ci] pre-commit autoupdate

updates:
- [github.com/adhtruong/mirrors-typos: v1.33.1 → v1.34.0](adhtruong/mirrors-typos@v1.33.1...v1.34.0)
- [github.com/astral-sh/ruff-pre-commit: v0.11.13 → v0.12.3](astral-sh/ruff-pre-commit@v0.11.13...v0.12.3)
- [github.com/woodruffw/zizmor-pre-commit: v1.9.0 → v1.11.0](zizmorcore/zizmor-pre-commit@v1.9.0...v1.11.0)
- [github.com/PyCQA/bandit: 1.8.3 → 1.8.6](PyCQA/bandit@1.8.3...1.8.6)

* Ignore B615

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Simon Alibert <[email protected]>
Co-authored-by: Simon Alibert <[email protected]>
@ben-z @pre-commit-ci @aliberts
Parse draccus subclass overrides when using --policy.path (huggingf…
1c0ac8e 
…ace#1501)

* Parse draccus subclass overrides when using --policy.path

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Simon Alibert <[email protected]>
@helper2424
Fix logging for mps in auto_select_torch_device (huggingface#1513)
3efb441
@michel-aractingi
Added missing licenses (huggingface#1517)
dfb1571 
* Added missing liscenses
@helper2424 @michel-aractingi
[Async Inference] Add gRPC retry mechanism to Async client (huggingfa…
8160349 
…ce#1485)

Co-authored-by: Michel Aractingi <[email protected]>
@pkooij
Feat/add bimanual so100 robot (huggingface#1509)
0938a1d
@imstevenpmwork @ben-z
Update pre-commit-config.yaml + pyproject.toml + ceil rerun & transfo…
378e1f0 
…rmer dependencies version (huggingface#1520)

* chore: update .gitignore

* chore: update pre-commit

* chore(deps): update pyproject

* fix(ci): multiple fixes

* chore: pre-commit apply

* chore: address review comments

* Update pyproject.toml

Co-authored-by: Ben Zhang <[email protected]>
Signed-off-by: Steven Palma <[email protected]>

* chore(deps): add todo

---------

Signed-off-by: Steven Palma <[email protected]>
Co-authored-by: Ben Zhang <[email protected]>
@jpizarrom
fix(hil-serl): drain queue on get_last_item_from_queue (huggingface#1524
7e9f955 
)

* fix(hil-serl): drain queue on get_last_item_from_queue

* parametrize queue tests

* revert changes for Darwin

* revert parametrize queue tests

* add test_get_last_item_multiple_items_with_torch_queue

* update test_get_last_item_multiple_items_with_torch_queue

* update test_get_last_item_multiple_items_with_torch_queue
@imstevenpmwork
feat(ci): add new & clean dockerfiles (huggingface#1525)
38d3737
@imstevenpmwork
chore(examples): remove outdated examples (huggingface#1526)
862a443
@zuoxingdong @CarolinePascal
Feat: Add Batched Video Encoding for Faster Dataset Recording (huggin…
e6e1f08 
…gface#1390)

* LeRobotDataset video encoding: updated `save_episode` method and added `batch_encode_videos` method to handle video encoding based on `batch_encoding_size`, allowing for both immediate and batched encoding.

* LeRobotDataset video cleanup: Enabled individual episode cleanup and check for remaining PNG files before removing the `images` directory.

* LeRobotDataset - VideoEncodingManager: added proper handling of pending episodes (encoding, cleaning) on exit or recording failures.

* LeRobotDatasetMetadata: removed `update_video_info` to only update video info at episode index 0 encoding.

* Adjusted the `record` function to utilize the new encoding management logic.

* Removed `encode_videos` method from `LeRobotDataset` and `encode_episode_videos` outputs as they are nowhere used.

---------

Signed-off-by: Xingdong Zuo <[email protected]>
Co-authored-by: Xingdong Zuo <[email protected]>
Co-authored-by: Caroline Pascal <[email protected]>
@AmanuelErgogo
merging with the upstream
1000bcd
@AmanuelErgogo AmanuelErgogo requested a review from umhan35 July 21, 2025 18:06
umhan35
umhan35 approved these changes Jul 23, 2025
View reviewed changes
Copy link

@umhan35 umhan35 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good job!

@umhan35 umhan35 changed the title Sync with Lerobot upstream Sync with Lerobot upstream for 101 use Jul 23, 2025
@umhan35 umhan35 merged commit 6749998 into main Jul 23, 2025
5 of 13 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@umhan35 umhan35 umhan35 approved these changes

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.