fix(tests) fixes for reachy2 tests; removing reachy2 references from the script

6 choose cameras in config

docs/source/_toctree.yml

@@ -41,8 +41,6 @@
 - sections:
   - local: notebooks
     title: Notebooks
-  - local: feetech
-    title: Updating Feetech Firmware
   title: "Resources"
 - sections:
   - local: contributing
@@ -50,7 +48,3 @@
   - local: backwardcomp
     title: Backward compatibility
   title: "About"
-- sections:
-  - local: datasets
-    title: "The LeRobotDataset Format"
-  -title: "Datasets"

docs/source/datasets.mdx

@@ -1,140 +0,0 @@
-# The LeRobotDataset Format
-
-`LeRobotDataset` is a standardized dataset format designed to address the specific needs of robot learning research.
-In this, it provides a unified and convenient access to robotics data across modalities, including sensorimotor readings, multiple camera feeds and teleoperation status.
-`LeRobotDataset` also stores general information regarding the data collected, like the task being performed by the teleoperator, the kind of robot used and measurement details like the frames per second at which the recording of both image and robot state's streams are proceeding.
-
-Therefore, `LeRobotDataset` provides a unified interface for handling multi-modal, time-series data, and it integrates seamlessly with the PyTorch and Hugging Face ecosystems.
-`LeRobotDataset` is designed to be easily extensible and customizable by users, and it already supports openly available data coming from a variety of embodiments, ranging from manipulator platforms like the SO-100 and ALOHA-2, to real-world humanoid data, simulation datasets and self-driving car datasets.
-This dataset format is built to be both efficient for training and flexible enough to accommodate the diverse data types encountered in robotics, while promoting reproducibility and ease of use for users.
-
-## The Format's Design
-
-A core design choice behind `LeRobotDataset` is separating the underlying data storage from the user-facing API.
-This allows for efficient serialization and storage while presenting the data in an intuitive, ready-to-use format.
-A dataset is always organized into three main components:
-
-1.  **Tabular Data**: Low-dimensional, high-frequency data such as joint states, and actions are stored in efficient [Apache Parquet](https://parquet.apache.org/) files, and typically offloaded to the more mature `datasets` library, providing fast, memory-mapped access.
-2.  **Visual Data**: To handle large volumes of camera data, frames are concatenated and encoded into MP4 files. Frames from the same episode are always grouped together into the same video, and multiple videos are grouped together by camera. To reduce stress on the file system, groups of videos for the same camera view are also broke into multiple sub-directories, after a given threshold number.
-3.  **Metadata**: A collection of JSON files which describes the dataset's structure in terms of its metadata, serving as the relational counterpart to both the tabular and visual dimensions of data. Metadata include the different feature schemas, frame rates, normalization statistics, and episode boundaries.
-
-For scalability, and to support datasets with potentially millions of trajectories resulting in hundreads of millions or billions of individual camera frames, we merge data from different episodes into the same high-level structure.
-Concretely, this means that any given tabular collection and video will not typically contain information about one episode only, but rather a concatenation of the information available in multiple episodes.
-This keeps the pressure on the file system, both locally and on remote storage providers like Hugging Face, manageable, at the expense of leveraging more heavily the metadata part of the data, e.g. used to reconstruct information relative to at which position a given episode starts or ends.
-An example structure for a given `LeRobotDataset` would appear as follows:
-
-```bash
-lerobot/svla_so101_pickplace
-├── data/
-│   └── chunk-000/
-│       ├── file_000000.parquet
-│       └── ...
-├── meta/
-│   ├── episodes/
-│   │   ├── chunk-000/
-│   │   │   └── file_000000.parquet
-│   │   └── ...
-│   ├── info.json
-│   ├── stats.json
-│   └── tasks.jsonl
-└── videos/
-    └── chunk-000/
-        ├── observation.images.wrist_camera/
-        │   ├── file_000000.mp4
-        │   └── ...
-        └── ...
-```
-
-- **`meta/info.json`**: This is the central metadata file. It contains the complete dataset schema, defining all features (e.g., `observation.state`, `action`), their shapes, and data types. It also stores crucial information like the dataset's frames-per-second (`fps`), codebase version, and the path templates used to locate data and video files.
-- **`meta/stats.json`**: This file stores aggregated statistics (mean, std, min, max) for each feature across the entire dataset. These are used for data normalization and are accessible via `dataset.meta.stats`.
-- **`meta/tasks.jsonl`**: Contains the mapping from natural language task descriptions to integer task indices, which are used for task-conditioned policy training.
-- **`meta/episodes/`**: This directory contains metadata about each individual episode, such as its length, corresponding task, and pointers to where its data is stored. For scalability, this information is stored in chunked Parquet files rather than a single large JSON file.
-- **`data/`**: Contains the core frame-by-frame tabular data in Parquet files. To improve performance and handle large datasets, data from **multiple episodes are concatenated into larger files**. These files are organized into chunked subdirectories to keep file sizes manageable. Therefore, a single file typically contains data for more than one episode.
-- **`videos/`**: Contains the MP4 video files for all visual observation streams. Similar to the `data/` directory, video footage from **multiple episodes is concatenated into single MP4 files**. This strategy significantly reduces the number of files in the dataset, which is more efficient for modern filesystems. The path structure (`/videos/<camera_key>/<chunk>/file_...mp4`) allows the data loader to locate the correct video file and then seek to the precise timestamp for a given frame.
-
-## Code Example: Using `LeRobotDataset` with `torch.utils.data.DataLoader`
-
-This section provides an overview of how to access datasets hosted on Hugging Face using the `LeRobotDataset` class.
-Every dataset on the Hugging Face Hub containing the three main pillars presented above (Tabular and Visual Data, as well as relational Metadata) can be assessed with a single line.
-Most reinforcement learning (RL) and behavioral cloning (BC) algorithms tend to operate on stack of observation and actions.
-For instance, RL algorithms typically use a history of previous observations `[o_{t-H}, ..., o_{t}]` to mitigate partial observability.
-BC cloning algorithms are instead typically trained to regress chunks of multiple actions rather than single controls.
-To accommodate for the specifics of robot learning training, `LeRobotDataset` provides a native windowing operation, whereby we can use the _seconds_ before and after any given observation using `delta_timestamps`.
-Non available frames is opportuninely padded, with a padding mask released to provide support in this.
-Notably, this all happens within the `LeRobotDataset` and is entitrely transparent to higher level wrappers such as `torch.utils.data.DataLoader`.
-
-Conveniently, by using `LeRobotDataset` with a Pytorch `DataLoader` one can automatically collate the individual sample dictionaries from the dataset into a single dictionary of batched tensors.
-
-```python
-from lerobot.datasets import LeRobotDataset
-
-# Load from the Hugging Face Hub (will be cached locally)
-dataset = LeRobotDataset("lerobot/svla_so101_pickplace")
-
-# Get the 100th frame in the dataset by
-sample = dataset[100]
-print(sample)
-# The sample is a dictionary of tensors
-# {
-#     'observation.state': tensor([...]),
-#     'action': tensor([...]),
-#     'observation.images.wrist_camera': tensor([C, H, W]),
-#     'timestamp': tensor(1.234),
-#     ...
-# }
-delta_timestamps = {
-    "observation.images.wrist_camera": [-0.2, -0.1, 0.0]  # 0.2, and 0.1 seconds *before* any observation
-}
-dataset = LeRobotDataset(
-    "lerobot/svla_so101_pickplace",
-    delta_timestamps=delta_timestamps
-)
-
-# Accessing an index now returns a stack of frames for the specified key
-sample = dataset[100]
-
-# The image tensor will now have a time dimension
-# 'observation.images.wrist_camera' has shape [T, C, H, W], where T=3
-print(sample['observation.images.wrist_camera'].shape)
-
-batch_size=16
-# wrap the dataset in a DataLoader to use process it batches for training purposes
-data_loader = torch.utils.data.DataLoader(
-    dataset,
-    batch_size=batch_size
-)
-
-# 3. Iterate over the DataLoader in a training loop
-num_epochs = 1
-device = "cuda" if torch.cuda.is_available() else "cpu"
-
-for epoch in range(num_epochs):
-    for batch in data_loader:
-        # 'batch' is a dictionary where each value is a batch of tensors.
-        # For example, batch['action'] will have a shape of [32, action_dim].
-
-        # If using delta_timestamps, a batched image tensor might have a
-        # shape of [32, T, C, H, W].
-
-        # Move data to the appropriate device (e.g., GPU)
-        observations = batch['observation.state'].to(device)
-        actions = batch['action'].to(device)
-        images = batch['observation.images.wrist_camera'].to(device)
-
-        # Next do amazing_model.forward(batch)
-        ...
-```
-
-## Streaming
-
-`LeRobotDataset` now also supports streaming mode.
-You can stream of data from a large dataset hosted on the Hugging Face Hub by just replacing the dataset definition with:
-
-```python
-from lerobot.datasets.streaming_dataset import StreamingLeRobotDataset
-
-# Streams frames from the Hugging Face Hub
-dataset = StreamingLeRobotDataset("lerobot/svla_so101_pickplace")
-```
-
-Streaming datasets supports high-performance batch processing (ca. 80-100 it/s, varying on connectivity) and high levels of frames randomization: a key feature for behavioral cloning algorithms otherwise operating on highly non-i.i.d. data.

docs/source/feetech.mdx

@@ -1,71 +0,0 @@
-# Feetech Motor Firmware Update
-
-This tutorial guides you through updating the firmware of Feetech motors using the official Feetech software.
-
-## Prerequisites
-
-- Windows computer (Feetech software is only available for Windows)
-- Feetech motor control board
-- USB cable to connect the control board to your computer
-- Feetech motors connected to the control board
-
-## Step 1: Download Feetech Software
-
-1. Visit the official Feetech software download page: [https://www.feetechrc.com/software.html](https://www.feetechrc.com/software.html)
-2. Download the latest version of the Feetech debugging software (FD)
-3. Install the software on your Windows computer
-
-## Step 2: Hardware Setup
-
-1. Connect your Feetech motors to the motor control board
-2. Connect the motor control board to your Windows computer via USB cable
-3. Ensure power is supplied to the motors
-
-## Step 3: Configure Connection
-
-1. Launch the Feetech debugging software
-2. Select the correct COM port from the port dropdown menu
-   - If unsure which port to use, check Windows Device Manager under "Ports (COM & LPT)"
-3. Set the appropriate baud rate (typically 1000000 for most Feetech motors)
-4. Click "Open" to establish communication with the control board
-
-## Step 4: Scan for Motors
-
-1. Once connected, click the "Search" button to detect all connected motors
-2. The software will automatically discover and list all motors on the bus
-3. Each motor will appear with its ID number
-
-## Step 5: Update Firmware
-
-For each motor you want to update:
-
-1. **Select the motor** from the list by clicking on it
-2. **Click on Upgrade tab**:
-3. **Click on Online button**:
-   - If an potential firmware update is found, it will be displayed in the box
-4. **Click on Upgrade button**:
-   - The update progress will be displayed
-
-## Step 6: Verify Update
-
-1. After the update completes, the software should automatically refresh the motor information
-2. Verify that the firmware version has been updated to the expected version
-
-## Important Notes
-
-⚠️ **Warning**: Do not disconnect power or USB during firmware updates, it will potentially brick the motor.
-
-## Bonus: Motor Debugging on Linux/macOS
-
-For debugging purposes only, you can use the open-source Feetech Debug Tool:
-
-- **Repository**: [FT_SCServo_Debug_Qt](https://github.com/CarolinePascal/FT_SCServo_Debug_Qt/tree/fix/port-search-timer)
-
-### Installation Instructions
-
-Follow the instructions in the repository to install the tool, for Ubuntu you can directly install it, for MacOS you need to build it from source.
-
-**Limitations:**
-
-- This tool is for debugging and parameter adjustment only
-- Firmware updates must still be done on Windows with official Feetech software

src/lerobot/datasets/lerobot_dataset.py

@@ -825,8 +825,6 @@ class LeRobotDataset(torch.utils.data.Dataset):
         """
         if not episode_data:
             episode_buffer = self.episode_buffer
-        else:
-            episode_buffer = episode_data
 
         validate_episode_buffer(episode_buffer, self.meta.total_episodes, self.features)
 

src/lerobot/record.py

@@ -209,14 +209,7 @@ def record_loop(
             (
                 t
                 for t in teleop
-                if isinstance(
-                    t,
-                    (
-                        so100_leader.SO100Leader,
-                        so101_leader.SO101Leader,
-                        koch_leader.KochLeader,
-                    ),
-                )
+                if isinstance(t, (so100_leader.SO100Leader, so101_leader.SO101Leader, koch_leader.KochLeader))
             ),
             None,
         )

src/lerobot/robots/bi_so100_follower/config_bi_so100_follower.py

@@ -29,10 +29,10 @@ class BiSO100FollowerConfig(RobotConfig):
 
     # Optional
     left_arm_disable_torque_on_disconnect: bool = True
-    left_arm_max_relative_target: float | dict[str, float] | None = None
+    left_arm_max_relative_target: int | None = None
     left_arm_use_degrees: bool = False
     right_arm_disable_torque_on_disconnect: bool = True
-    right_arm_max_relative_target: float | dict[str, float] | None = None
+    right_arm_max_relative_target: int | None = None
     right_arm_use_degrees: bool = False
 
     # cameras (shared between both arms)

src/lerobot/robots/hope_jr/config_hope_jr.py

@@ -44,8 +44,8 @@ class HopeJrArmConfig(RobotConfig):
     disable_torque_on_disconnect: bool = True
 
     # `max_relative_target` limits the magnitude of the relative positional target vector for safety purposes.
-    # Set this to a positive scalar to have the same value for all motors, or a dictionary that maps motor
-    # names to the max_relative_target value for that motor.
-    max_relative_target: float | dict[str, float] | None = None
+    # Set this to a positive scalar to have the same value for all motors, or a list that is the same length as
+    # the number of motors in your follower arms.
+    max_relative_target: int | None = None
 
     cameras: dict[str, CameraConfig] = field(default_factory=dict)

src/lerobot/robots/koch_follower/config_koch_follower.py

@@ -28,9 +28,9 @@ class KochFollowerConfig(RobotConfig):
     disable_torque_on_disconnect: bool = True
 
     # `max_relative_target` limits the magnitude of the relative positional target vector for safety purposes.
-    # Set this to a positive scalar to have the same value for all motors, or a dictionary that maps motor
-    # names to the max_relative_target value for that motor.
-    max_relative_target: float | dict[str, float] | None = None
+    # Set this to a positive scalar to have the same value for all motors, or a list that is the same length as
+    # the number of motors in your follower arms.
+    max_relative_target: int | None = None
 
     # cameras
     cameras: dict[str, CameraConfig] = field(default_factory=dict)

src/lerobot/robots/koch_follower/koch_follower.py

@@ -110,7 +110,6 @@ class KochFollower(Robot):
         return self.bus.is_calibrated
 
     def calibrate(self) -> None:
-        self.bus.disable_torque()
         if self.calibration:
             # Calibration file exists, ask user whether to use it or run new calibration
             user_input = input(
@@ -121,6 +120,7 @@ class KochFollower(Robot):
                 self.bus.write_calibration(self.calibration)
                 return
         logger.info(f"\nRunning calibration of {self}")
+        self.bus.disable_torque()
         for motor in self.bus.motors:
             self.bus.write("Operating_Mode", motor, OperatingMode.EXTENDED_POSITION.value)
 

src/lerobot/robots/lekiwi/config_lekiwi.py

@@ -39,9 +39,9 @@ class LeKiwiConfig(RobotConfig):
     disable_torque_on_disconnect: bool = True
 
     # `max_relative_target` limits the magnitude of the relative positional target vector for safety purposes.
-    # Set this to a positive scalar to have the same value for all motors, or a dictionary that maps motor
-    # names to the max_relative_target value for that motor.
-    max_relative_target: float | dict[str, float] | None = None
+    # Set this to a positive scalar to have the same value for all motors, or a list that is the same length as
+    # the number of motors in your follower arms.
+    max_relative_target: int | None = None
 
     cameras: dict[str, CameraConfig] = field(default_factory=lekiwi_cameras_config)
 

src/lerobot/robots/so100_follower/config_so100_follower.py

@@ -30,9 +30,9 @@ class SO100FollowerConfig(RobotConfig):
     disable_torque_on_disconnect: bool = True
 
     # `max_relative_target` limits the magnitude of the relative positional target vector for safety purposes.
-    # Set this to a positive scalar to have the same value for all motors, or a dictionary that maps motor
-    # names to the max_relative_target value for that motor.
-    max_relative_target: float | dict[str, float] | None = None
+    # Set this to a positive scalar to have the same value for all motors, or a list that is the same length as
+    # the number of motors in your follower arms.
+    max_relative_target: int | None = None
 
     # cameras
     cameras: dict[str, CameraConfig] = field(default_factory=dict)

src/lerobot/robots/so100_follower/so100_follower.py

@@ -161,11 +161,6 @@ class SO100Follower(Robot):
                 self.bus.write("I_Coefficient", motor, 0)
                 self.bus.write("D_Coefficient", motor, 32)
 
-                if motor == "gripper":
-                    self.bus.write("Max_Torque_Limit", motor, 500)  # 50% of max torque to avoid burnout
-                    self.bus.write("Protection_Current", motor, 250)  # 50% of max current to avoid burnout
-                    self.bus.write("Overload_Torque", motor, 25)  # 25% torque when overloaded
-
     def setup_motors(self) -> None:
         for motor in reversed(self.bus.motors):
             input(f"Connect the controller board to the '{motor}' motor only and press enter.")

src/lerobot/robots/so101_follower/config_so101_follower.py

@@ -30,9 +30,9 @@ class SO101FollowerConfig(RobotConfig):
     disable_torque_on_disconnect: bool = True
 
     # `max_relative_target` limits the magnitude of the relative positional target vector for safety purposes.
-    # Set this to a positive scalar to have the same value for all motors, or a dictionary that maps motor
-    # names to the max_relative_target value for that motor.
-    max_relative_target: float | dict[str, float] | None = None
+    # Set this to a positive scalar to have the same value for all motors, or a list that is the same length as
+    # the number of motors in your follower arms.
+    max_relative_target: int | None = None
 
     # cameras
     cameras: dict[str, CameraConfig] = field(default_factory=dict)

src/lerobot/robots/so101_follower/so101_follower.py

@@ -157,13 +157,6 @@ class SO101Follower(Robot):
                 self.bus.write("I_Coefficient", motor, 0)
                 self.bus.write("D_Coefficient", motor, 32)
 
-                if motor == "gripper":
-                    self.bus.write(
-                        "Max_Torque_Limit", motor, 500
-                    )  # 50% of the max torque limit to avoid burnout
-                    self.bus.write("Protection_Current", motor, 250)  # 50% of max current to avoid burnout
-                    self.bus.write("Overload_Torque", motor, 25)  # 25% torque when overloaded
-
     def setup_motors(self) -> None:
         for motor in reversed(self.bus.motors):
             input(f"Connect the controller board to the '{motor}' motor only and press enter.")

src/lerobot/robots/stretch3/configuration_stretch3.py

@@ -24,6 +24,11 @@ from ..config import RobotConfig
 @RobotConfig.register_subclass("stretch3")
 @dataclass
 class Stretch3RobotConfig(RobotConfig):
+    # `max_relative_target` limits the magnitude of the relative positional target vector for safety purposes.
+    # Set this to a positive scalar to have the same value for all motors, or a list that is the same length as
+    # the number of motors in your follower arms.
+    max_relative_target: int | None = None
+
     # cameras
     cameras: dict[str, CameraConfig] = field(
         default_factory=lambda: {

src/lerobot/robots/utils.py

@@ -74,7 +74,7 @@ def make_robot_from_config(config: RobotConfig) -> Robot:
 
 
 def ensure_safe_goal_position(
-    goal_present_pos: dict[str, tuple[float, float]], max_relative_target: float | dict[str, float]
+    goal_present_pos: dict[str, tuple[float, float]], max_relative_target: float | dict[float]
 ) -> dict[str, float]:
     """Caps relative action target magnitude for safety."""
 

src/lerobot/robots/viperx/config_viperx.py

@@ -28,15 +28,15 @@ class ViperXConfig(RobotConfig):
 
     # /!\ FOR SAFETY, READ THIS /!\
     # `max_relative_target` limits the magnitude of the relative positional target vector for safety purposes.
-    # Set this to a positive scalar to have the same value for all motors, or a dictionary that maps motor
-    # names to the max_relative_target value for that motor.
+    # Set this to a positive scalar to have the same value for all motors, or a list that is the same length as
+    # the number of motors in your follower arms.
     # For Aloha, for every goal position request, motor rotations are capped at 5 degrees by default.
     # When you feel more confident with teleoperation or running the policy, you can extend
     # this safety limit and even removing it by setting it to `null`.
     # Also, everything is expected to work safely out-of-the-box, but we highly advise to
     # first try to teleoperate the grippers only (by commenting out the rest of the motors in this yaml),
     # then to gradually add more motors (by uncommenting), until you can teleoperate both arms fully
-    max_relative_target: float | dict[str, float] = 5.0
+    max_relative_target: int | None = 5
 
     # cameras
     cameras: dict[str, CameraConfig] = field(default_factory=dict)

src/lerobot/scripts/server/robot_client.py

@@ -302,6 +302,11 @@ class RobotClient:
 
                     self.logger.debug(f"Current latest action: {latest_action}")
 
+                    # Get queue state before changes
+                    old_size, old_timesteps = self._inspect_action_queue()
+                    if not old_timesteps:
+                        old_timesteps = [latest_action]  # queue was empty
+
                     # Get queue state before changes
                     old_size, old_timesteps = self._inspect_action_queue()
                     if not old_timesteps:

src/lerobot/teleoperators/koch_leader/koch_leader.py

@@ -88,7 +88,6 @@ class KochLeader(Teleoperator):
         return self.bus.is_calibrated
 
     def calibrate(self) -> None:
-        self.bus.disable_torque()
         if self.calibration:
             # Calibration file exists, ask user whether to use it or run new calibration
             user_input = input(
@@ -99,6 +98,7 @@ class KochLeader(Teleoperator):
                 self.bus.write_calibration(self.calibration)
                 return
         logger.info(f"\nRunning calibration of {self}")
+        self.bus.disable_torque()
         for motor in self.bus.motors:
             self.bus.write("Operating_Mode", motor, OperatingMode.EXTENDED_POSITION.value)
 

tests/cameras/test_reachy2_camera.py

@@ -23,6 +23,8 @@ import pytest
 from lerobot.cameras.reachy2_camera import Reachy2Camera, Reachy2CameraConfig
 from lerobot.errors import DeviceNotConnectedError
 
+pytest.importorskip("reachy2_sdk")
+
 PARAMS = [
     ("teleop", "left"),
     ("teleop", "right"),

tests/conftest.py

@@ -28,7 +28,6 @@ pytest_plugins = [
     "tests.fixtures.files",
     "tests.fixtures.hub",
     "tests.fixtures.optimizers",
-    "tests.plugins.reachy2_sdk",
 ]
 
 

tests/plugins/reachy2_sdk.py

@@ -1,30 +0,0 @@
-import sys
-import types
-from unittest.mock import MagicMock
-
-
-def _install_reachy2_sdk_stub():
-    sdk = types.ModuleType("reachy2_sdk")
-    sdk.__path__ = []
-    sdk.ReachySDK = MagicMock(name="ReachySDK")
-
-    media = types.ModuleType("reachy2_sdk.media")
-    media.__path__ = []
-    camera = types.ModuleType("reachy2_sdk.media.camera")
-    camera.CameraView = MagicMock(name="CameraView")
-    camera_manager = types.ModuleType("reachy2_sdk.media.camera_manager")
-    camera_manager.CameraManager = MagicMock(name="CameraManager")
-
-    sdk.media = media
-    media.camera = camera
-    media.camera_manager = camera_manager
-
-    # Register in sys.modules
-    sys.modules.setdefault("reachy2_sdk", sdk)
-    sys.modules.setdefault("reachy2_sdk.media", media)
-    sys.modules.setdefault("reachy2_sdk.media.camera", camera)
-    sys.modules.setdefault("reachy2_sdk.media.camera_manager", camera_manager)
-
-
-def pytest_sessionstart(session):
-    _install_reachy2_sdk_stub()

tests/robots/test_reachy2.py

@@ -51,6 +51,8 @@ PARAMS = [
     {"with_left_teleop_camera": False, "with_torso_camera": True},
 ]
 
+pytest.importorskip("reachy2_sdk")
+
 
 def _make_reachy2_sdk_mock():
     class JointSpy:

tests/teleoperators/test_reachy2_teleoperator.py

@@ -28,6 +28,8 @@ from lerobot.teleoperators.reachy2_teleoperator import (
     Reachy2TeleoperatorConfig,
 )
 
+pytest.importorskip("reachy2_sdk")
+
 # {lerobot_keys: reachy2_sdk_keys}
 REACHY2_JOINTS = {
     **REACHY2_NECK_JOINTS,