Add FeetechMotorsBus, SO-100, Moss-v1 (#419)

Co-authored-by: jess-moss <jess.moss@huggingface.co>
Co-authored-by: Simon Alibert <75076266+aliberts@users.noreply.github.com>

lerobot/configs/env/moss_real.yaml

@@ -0,0 +1,10 @@
+# @package _global_
+
+fps: 30
+
+env:
+  name: real_world
+  task: null
+  state_dim: 6
+  action_dim: 6
+  fps: ${fps}

lerobot/configs/env/so100_real.yaml

@@ -0,0 +1,10 @@
+# @package _global_
+
+fps: 30
+
+env:
+  name: real_world
+  task: null
+  state_dim: 6
+  action_dim: 6
+  fps: ${fps}

lerobot/configs/policy/act_moss_real.yaml

@@ -0,0 +1,102 @@
+# @package _global_
+
+# Use `act_koch_real.yaml` to train on real-world datasets collected on Alexander Koch's robots.
+# Compared to `act.yaml`, it contains 2 cameras (i.e. laptop, phone) instead of 1 camera (i.e. top).
+# Also, `training.eval_freq` is set to -1. This config is used to evaluate checkpoints at a certain frequency of training steps.
+# When it is set to -1, it deactivates evaluation. This is because real-world evaluation is done through our `control_robot.py` script.
+# Look at the documentation in header of `control_robot.py` for more information on how to collect data , train and evaluate a policy.
+#
+# Example of usage for training:
+# ```bash
+# python lerobot/scripts/train.py \
+#   policy=act_koch_real \
+#   env=koch_real
+# ```
+
+seed: 1000
+dataset_repo_id: lerobot/moss_pick_place_lego
+
+override_dataset_stats:
+  observation.images.laptop:
+    # stats from imagenet, since we use a pretrained vision model
+    mean: [[[0.485]], [[0.456]], [[0.406]]]  # (c,1,1)
+    std: [[[0.229]], [[0.224]], [[0.225]]]  # (c,1,1)
+  observation.images.phone:
+    # stats from imagenet, since we use a pretrained vision model
+    mean: [[[0.485]], [[0.456]], [[0.406]]]  # (c,1,1)
+    std: [[[0.229]], [[0.224]], [[0.225]]]  # (c,1,1)
+
+training:
+  offline_steps: 80000
+  online_steps: 0
+  eval_freq: -1
+  save_freq: 10000
+  log_freq: 100
+  save_checkpoint: true
+
+  batch_size: 8
+  lr: 1e-5
+  lr_backbone: 1e-5
+  weight_decay: 1e-4
+  grad_clip_norm: 10
+  online_steps_between_rollouts: 1
+
+  delta_timestamps:
+    action: "[i / ${fps} for i in range(${policy.chunk_size})]"
+
+eval:
+  n_episodes: 50
+  batch_size: 50
+
+# See `configuration_act.py` for more details.
+policy:
+  name: act
+
+  # Input / output structure.
+  n_obs_steps: 1
+  chunk_size: 100
+  n_action_steps: 100
+
+  input_shapes:
+    # TODO(rcadene, alexander-soare): add variables for height and width from the dataset/env?
+    observation.images.laptop: [3, 480, 640]
+    observation.images.phone: [3, 480, 640]
+    observation.state: ["${env.state_dim}"]
+  output_shapes:
+    action: ["${env.action_dim}"]
+
+  # Normalization / Unnormalization
+  input_normalization_modes:
+    observation.images.laptop: mean_std
+    observation.images.phone: mean_std
+    observation.state: mean_std
+  output_normalization_modes:
+    action: mean_std
+
+  # Architecture.
+  # Vision backbone.
+  vision_backbone: resnet18
+  pretrained_backbone_weights: ResNet18_Weights.IMAGENET1K_V1
+  replace_final_stride_with_dilation: false
+  # Transformer layers.
+  pre_norm: false
+  dim_model: 512
+  n_heads: 8
+  dim_feedforward: 3200
+  feedforward_activation: relu
+  n_encoder_layers: 4
+  # Note: Although the original ACT implementation has 7 for `n_decoder_layers`, there is a bug in the code
+  # that means only the first layer is used. Here we match the original implementation by setting this to 1.
+  # See this issue https://github.com/tonyzhaozh/act/issues/25#issue-2258740521.
+  n_decoder_layers: 1
+  # VAE.
+  use_vae: true
+  latent_dim: 32
+  n_vae_encoder_layers: 4
+
+  # Inference.
+  temporal_ensemble_momentum: null
+
+  # Training and loss computation.
+  dropout: 0.1
+  kl_weight: 10.0

lerobot/configs/policy/act_so100_real.yaml

@@ -0,0 +1,102 @@
+# @package _global_
+
+# Use `act_koch_real.yaml` to train on real-world datasets collected on Alexander Koch's robots.
+# Compared to `act.yaml`, it contains 2 cameras (i.e. laptop, phone) instead of 1 camera (i.e. top).
+# Also, `training.eval_freq` is set to -1. This config is used to evaluate checkpoints at a certain frequency of training steps.
+# When it is set to -1, it deactivates evaluation. This is because real-world evaluation is done through our `control_robot.py` script.
+# Look at the documentation in header of `control_robot.py` for more information on how to collect data , train and evaluate a policy.
+#
+# Example of usage for training:
+# ```bash
+# python lerobot/scripts/train.py \
+#   policy=act_koch_real \
+#   env=koch_real
+# ```
+
+seed: 1000
+dataset_repo_id: lerobot/so100_pick_place_lego
+
+override_dataset_stats:
+  observation.images.laptop:
+    # stats from imagenet, since we use a pretrained vision model
+    mean: [[[0.485]], [[0.456]], [[0.406]]]  # (c,1,1)
+    std: [[[0.229]], [[0.224]], [[0.225]]]  # (c,1,1)
+  observation.images.phone:
+    # stats from imagenet, since we use a pretrained vision model
+    mean: [[[0.485]], [[0.456]], [[0.406]]]  # (c,1,1)
+    std: [[[0.229]], [[0.224]], [[0.225]]]  # (c,1,1)
+
+training:
+  offline_steps: 80000
+  online_steps: 0
+  eval_freq: -1
+  save_freq: 10000
+  log_freq: 100
+  save_checkpoint: true
+
+  batch_size: 8
+  lr: 1e-5
+  lr_backbone: 1e-5
+  weight_decay: 1e-4
+  grad_clip_norm: 10
+  online_steps_between_rollouts: 1
+
+  delta_timestamps:
+    action: "[i / ${fps} for i in range(${policy.chunk_size})]"
+
+eval:
+  n_episodes: 50
+  batch_size: 50
+
+# See `configuration_act.py` for more details.
+policy:
+  name: act
+
+  # Input / output structure.
+  n_obs_steps: 1
+  chunk_size: 100
+  n_action_steps: 100
+
+  input_shapes:
+    # TODO(rcadene, alexander-soare): add variables for height and width from the dataset/env?
+    observation.images.laptop: [3, 480, 640]
+    observation.images.phone: [3, 480, 640]
+    observation.state: ["${env.state_dim}"]
+  output_shapes:
+    action: ["${env.action_dim}"]
+
+  # Normalization / Unnormalization
+  input_normalization_modes:
+    observation.images.laptop: mean_std
+    observation.images.phone: mean_std
+    observation.state: mean_std
+  output_normalization_modes:
+    action: mean_std
+
+  # Architecture.
+  # Vision backbone.
+  vision_backbone: resnet18
+  pretrained_backbone_weights: ResNet18_Weights.IMAGENET1K_V1
+  replace_final_stride_with_dilation: false
+  # Transformer layers.
+  pre_norm: false
+  dim_model: 512
+  n_heads: 8
+  dim_feedforward: 3200
+  feedforward_activation: relu
+  n_encoder_layers: 4
+  # Note: Although the original ACT implementation has 7 for `n_decoder_layers`, there is a bug in the code
+  # that means only the first layer is used. Here we match the original implementation by setting this to 1.
+  # See this issue https://github.com/tonyzhaozh/act/issues/25#issue-2258740521.
+  n_decoder_layers: 1
+  # VAE.
+  use_vae: true
+  latent_dim: 32
+  n_vae_encoder_layers: 4
+
+  # Inference.
+  temporal_ensemble_momentum: null
+
+  # Training and loss computation.
+  dropout: 0.1
+  kl_weight: 10.0

lerobot/configs/robot/aloha.yaml

@@ -1,11 +1,13 @@
-# Aloha: A Low-Cost Hardware for Bimanual Teleoperation
+# [Aloha: A Low-Cost Hardware for Bimanual Teleoperation](https://www.trossenrobotics.com/aloha-stationary)
 # https://aloha-2.github.io
-# https://www.trossenrobotics.com/aloha-stationary
 
 # Requires installing extras packages
 # With pip: `pip install -e ".[dynamixel intelrealsense]"`
 # With poetry: `poetry install --sync --extras "dynamixel intelrealsense"`
 
+# See [tutorial](https://github.com/huggingface/lerobot/blob/main/examples/9_use_aloha.md)
+
+
 _target_: lerobot.common.robot_devices.robots.manipulator.ManipulatorRobot
 robot_type: aloha
 # Specific to Aloha, LeRobot comes with default calibration files. Assuming the motors have been

lerobot/configs/robot/koch_bimanual.yaml

@@ -1,5 +1,5 @@
 _target_: lerobot.common.robot_devices.robots.manipulator.ManipulatorRobot
-robot_type: koch
+robot_type: koch_bimanual
 calibration_dir: .cache/calibration/koch_bimanual
 
 # `max_relative_target` limits the magnitude of the relative positional target vector for safety purposes.

lerobot/configs/robot/moss.yaml

@@ -0,0 +1,56 @@
+# [Moss v1 robot arm](https://github.com/jess-moss/moss-robot-arms)
+
+# Requires installing extras packages
+# With pip: `pip install -e ".[feetech]"`
+# With poetry: `poetry install --sync --extras "feetech"`
+
+# See [tutorial](https://github.com/huggingface/lerobot/blob/main/examples/11_use_moss.md)
+
+_target_: lerobot.common.robot_devices.robots.manipulator.ManipulatorRobot
+robot_type: moss
+calibration_dir: .cache/calibration/moss
+
+# `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: null
+
+leader_arms:
+  main:
+    _target_: lerobot.common.robot_devices.motors.feetech.FeetechMotorsBus
+    port: /dev/tty.usbmodem58760431091
+    motors:
+      # name: (index, model)
+      shoulder_pan: [1, "sts3215"]
+      shoulder_lift: [2, "sts3215"]
+      elbow_flex: [3, "sts3215"]
+      wrist_flex: [4, "sts3215"]
+      wrist_roll: [5, "sts3215"]
+      gripper: [6, "sts3215"]
+
+follower_arms:
+  main:
+    _target_: lerobot.common.robot_devices.motors.feetech.FeetechMotorsBus
+    port: /dev/tty.usbmodem58760431191
+    motors:
+      # name: (index, model)
+      shoulder_pan: [1, "sts3215"]
+      shoulder_lift: [2, "sts3215"]
+      elbow_flex: [3, "sts3215"]
+      wrist_flex: [4, "sts3215"]
+      wrist_roll: [5, "sts3215"]
+      gripper: [6, "sts3215"]
+
+cameras:
+  laptop:
+    _target_: lerobot.common.robot_devices.cameras.opencv.OpenCVCamera
+    camera_index: 0
+    fps: 30
+    width: 640
+    height: 480
+  phone:
+    _target_: lerobot.common.robot_devices.cameras.opencv.OpenCVCamera
+    camera_index: 1
+    fps: 30
+    width: 640
+    height: 480

lerobot/configs/robot/so100.yaml

@@ -0,0 +1,56 @@
+# [SO-100 robot arm](https://github.com/TheRobotStudio/SO-ARM100)
+
+# Requires installing extras packages
+# With pip: `pip install -e ".[feetech]"`
+# With poetry: `poetry install --sync --extras "feetech"`
+
+# See [tutorial](https://github.com/huggingface/lerobot/blob/main/examples/10_use_so100.md)
+
+_target_: lerobot.common.robot_devices.robots.manipulator.ManipulatorRobot
+robot_type: so100
+calibration_dir: .cache/calibration/so100
+
+# `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: null
+
+leader_arms:
+  main:
+    _target_: lerobot.common.robot_devices.motors.feetech.FeetechMotorsBus
+    port: /dev/tty.usbmodem585A0077581
+    motors:
+      # name: (index, model)
+      shoulder_pan: [1, "sts3215"]
+      shoulder_lift: [2, "sts3215"]
+      elbow_flex: [3, "sts3215"]
+      wrist_flex: [4, "sts3215"]
+      wrist_roll: [5, "sts3215"]
+      gripper: [6, "sts3215"]
+
+follower_arms:
+  main:
+    _target_: lerobot.common.robot_devices.motors.feetech.FeetechMotorsBus
+    port: /dev/tty.usbmodem585A0080971
+    motors:
+      # name: (index, model)
+      shoulder_pan: [1, "sts3215"]
+      shoulder_lift: [2, "sts3215"]
+      elbow_flex: [3, "sts3215"]
+      wrist_flex: [4, "sts3215"]
+      wrist_roll: [5, "sts3215"]
+      gripper: [6, "sts3215"]
+
+cameras:
+  laptop:
+    _target_: lerobot.common.robot_devices.cameras.opencv.OpenCVCamera
+    camera_index: 0
+    fps: 30
+    width: 640
+    height: 480
+  phone:
+    _target_: lerobot.common.robot_devices.cameras.opencv.OpenCVCamera
+    camera_index: 1
+    fps: 30
+    width: 640
+    height: 480

lerobot/configs/robot/stretch.yaml

@@ -1,3 +1,12 @@
+# [Stretch3 from Hello Robot](https://hello-robot.com/stretch-3-product)
+
+# Requires installing extras packages
+# With pip: `pip install -e ".[stretch]"`
+# With poetry: `poetry install --sync --extras "stretch"`
+
+# See [tutorial](https://github.com/huggingface/lerobot/blob/main/examples/8_use_stretch.md)
+
+
 _target_: lerobot.common.robot_devices.robots.stretch.StretchRobot
 robot_type: stretch3