resume: false
device: cuda
use_amp: false
seed: 1000
dataset_repo_id: kywch/mimicgen_stack_d1
video_backend: pyav
training:
  offline_steps: 100000
  num_workers: 4
  batch_size: 8
  eval_freq: 10000
  log_freq: 200
  save_checkpoint: true
  save_freq: 20000
  online_steps: 0
  online_rollout_n_episodes: 1
  online_rollout_batch_size: 1
  online_steps_between_rollouts: 1
  online_sampling_ratio: 0.5
  online_env_seed: null
  online_buffer_capacity: null
  online_buffer_seed_size: 0
  do_online_rollout_async: false
  image_transforms:
    enable: false
    max_num_transforms: 3
    random_order: false
    brightness:
      weight: 1
      min_max:
      - 0.8
      - 1.2
    contrast:
      weight: 1
      min_max:
      - 0.8
      - 1.2
    saturation:
      weight: 1
      min_max:
      - 0.5
      - 1.5
    hue:
      weight: 1
      min_max:
      - -0.05
      - 0.05
    sharpness:
      weight: 1
      min_max:
      - 0.8
      - 1.2
  lr: 1.0e-05
  lr_backbone: 1.0e-05
  weight_decay: 0.0001
  grad_clip_norm: 10
  delta_timestamps:
    action:
    - 0.0
    - 0.05
    - 0.1
    - 0.15
    - 0.2
    - 0.25
    - 0.3
    - 0.35
    - 0.4
    - 0.45
    - 0.5
    - 0.55
    - 0.6
    - 0.65
    - 0.7
    - 0.75
    - 0.8
    - 0.85
    - 0.9
    - 0.95
eval:
  n_episodes: 20
  batch_size: 20
  use_async_envs: true
wandb:
  enable: true
  disable_artifact: true
  project: lerobot
  notes: ''
fps: 20
env:
  name: mimicgen
  task: Stack_D1
  state_dim: 9
  action_dim: 7
  episode_length: 250
  meta: stack_d1_env.json
  image_keys:
  - agentview
  - robot0_eye_in_hand
  state_keys:
  - robot0_eef_pos
  - robot0_eef_quat
  - robot0_gripper_qpos
  use_delta_action: false
  use_highres_image_obs: true
override_dataset_stats:
  observation.images.agentview:
    mean:
    - - - 0.485
    - - - 0.456
    - - - 0.406
    std:
    - - - 0.229
    - - - 0.224
    - - - 0.225
  observation.images.robot0_eye_in_hand:
    mean:
    - - - 0.485
    - - - 0.456
    - - - 0.406
    std:
    - - - 0.229
    - - - 0.224
    - - - 0.225
policy:
  name: act
  n_obs_steps: 1
  chunk_size: 20
  n_action_steps: 20
  input_shapes:
    observation.images.agentview:
    - 3
    - 256
    - 256
    observation.images.robot0_eye_in_hand:
    - 3
    - 256
    - 256
    observation.state:
    - ${env.state_dim}
  output_shapes:
    action:
    - ${env.action_dim}
  input_normalization_modes:
    observation.images.agentview: mean_std
    observation.images.robot0_eye_in_hand: mean_std
    observation.state: mean_std
  output_normalization_modes:
    action: mean_std
  vision_backbone: resnet18
  pretrained_backbone_weights: ResNet18_Weights.IMAGENET1K_V1
  replace_final_stride_with_dilation: false
  pre_norm: false
  dim_model: 512
  n_heads: 8
  dim_feedforward: 3200
  feedforward_activation: relu
  n_encoder_layers: 4
  n_decoder_layers: 1
  use_vae: true
  latent_dim: 32
  n_vae_encoder_layers: 4
  temporal_ensemble_momentum: null
  dropout: 0.1
  kl_weight: 10.0