Add files using upload-large-folder tool

artifacts/generated_configs/proxy_interaction_r3d_stage1_clip_seed7.yaml

@@ -0,0 +1,149 @@
+experiment_name: proxy_interaction_r3d_stage1_clip_seed7
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 7
+init_checkpoint: /workspace/VLAarchtests/artifacts/outputs/reveal_runs/proxy_backbone_only_clip/checkpoint_best.pt
+init_strict: false
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 224
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_clip224_v6_rgbd_stage1_seed7.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_clip224_v6_rgbd_stage1_seed7.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 7
+optim:
+  epochs: 4
+  batch_size: 2
+  num_workers: 4
+  lr: 0.0003
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: true
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 512
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: false
+  fusion:
+    hidden_dim: 512
+    num_cameras: 3
+    num_layers: 4
+    num_heads: 8
+    ff_dim: 2048
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 512
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 8
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 512
+    num_heads: 8
+    num_layers: 4
+    ff_dim: 2048
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 512
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 8
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 512
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 512
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 8
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.1
+  arm_role: 0.15
+  support_mode: 0.1
+  corridor: 0.15
+  persistence: 0.05
+  disturbance: 0.05
+  world_model: 0.2
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.25
+  planner_risk: 0.1
+  planner_ranking: 0.2
+  proposal_reconstruction: 0.1
+  proposal_success: 0.15
+  proposal_ranking: 0.2
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage1_clip_seed8.yaml

@@ -0,0 +1,149 @@
+experiment_name: proxy_interaction_r3d_stage1_clip_seed8
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 8
+init_checkpoint: /workspace/VLAarchtests/artifacts/outputs/reveal_runs/proxy_backbone_only_clip/checkpoint_best.pt
+init_strict: false
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 224
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_clip224_v6_rgbd_stage1_seed8.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_clip224_v6_rgbd_stage1_seed8.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 8
+optim:
+  epochs: 4
+  batch_size: 2
+  num_workers: 4
+  lr: 0.0003
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: true
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 512
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: false
+  fusion:
+    hidden_dim: 512
+    num_cameras: 3
+    num_layers: 4
+    num_heads: 8
+    ff_dim: 2048
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 512
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 8
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 512
+    num_heads: 8
+    num_layers: 4
+    ff_dim: 2048
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 512
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 8
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 512
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 512
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 8
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.1
+  arm_role: 0.15
+  support_mode: 0.1
+  corridor: 0.15
+  persistence: 0.05
+  disturbance: 0.05
+  world_model: 0.2
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.25
+  planner_risk: 0.1
+  planner_ranking: 0.2
+  proposal_reconstruction: 0.1
+  proposal_success: 0.15
+  proposal_ranking: 0.2
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage1_clip_seed9.yaml

@@ -0,0 +1,149 @@
+experiment_name: proxy_interaction_r3d_stage1_clip_seed9
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 9
+init_checkpoint: /workspace/VLAarchtests/artifacts/outputs/reveal_runs/proxy_backbone_only_clip/checkpoint_best.pt
+init_strict: false
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 224
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_clip224_v6_rgbd_stage1_seed9.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_clip224_v6_rgbd_stage1_seed9.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 9
+optim:
+  epochs: 4
+  batch_size: 2
+  num_workers: 4
+  lr: 0.0003
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: true
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 512
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: false
+  fusion:
+    hidden_dim: 512
+    num_cameras: 3
+    num_layers: 4
+    num_heads: 8
+    ff_dim: 2048
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 512
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 8
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 512
+    num_heads: 8
+    num_layers: 4
+    ff_dim: 2048
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 512
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 8
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 512
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 512
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 8
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.1
+  arm_role: 0.15
+  support_mode: 0.1
+  corridor: 0.15
+  persistence: 0.05
+  disturbance: 0.05
+  world_model: 0.2
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.25
+  planner_risk: 0.1
+  planner_ranking: 0.2
+  proposal_reconstruction: 0.1
+  proposal_success: 0.15
+  proposal_ranking: 0.2
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage1_dummy_seed13.yaml

@@ -0,0 +1,147 @@
+experiment_name: proxy_interaction_r3d_stage1_dummy_seed13
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 13
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 96
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_v6_rgbd_stage1_dummy_seed13.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_v6_rgbd_stage1_dummy_seed13.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 13
+optim:
+  epochs: 4
+  batch_size: 16
+  num_workers: 4
+  lr: 0.001
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: false
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 192
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: true
+  fusion:
+    hidden_dim: 192
+    num_cameras: 3
+    num_layers: 2
+    num_heads: 4
+    ff_dim: 384
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 192
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 4
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 192
+    num_heads: 4
+    num_layers: 2
+    ff_dim: 384
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 192
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 4
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 192
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 4
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 192
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 4
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.15
+  arm_role: 0.2
+  support_mode: 0.15
+  corridor: 0.2
+  persistence: 0.1
+  disturbance: 0.1
+  world_model: 0.25
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.2
+  planner_risk: 0.1
+  planner_ranking: 0.1
+  proposal_reconstruction: 0.2
+  proposal_success: 0.1
+  proposal_ranking: 0.1
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage1_dummy_seed14.yaml

@@ -0,0 +1,147 @@
+experiment_name: proxy_interaction_r3d_stage1_dummy_seed14
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 14
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 96
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_v6_rgbd_stage1_dummy_seed14.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_v6_rgbd_stage1_dummy_seed14.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 14
+optim:
+  epochs: 4
+  batch_size: 16
+  num_workers: 4
+  lr: 0.001
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: false
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 192
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: true
+  fusion:
+    hidden_dim: 192
+    num_cameras: 3
+    num_layers: 2
+    num_heads: 4
+    ff_dim: 384
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 192
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 4
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 192
+    num_heads: 4
+    num_layers: 2
+    ff_dim: 384
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 192
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 4
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 192
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 4
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 192
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 4
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.15
+  arm_role: 0.2
+  support_mode: 0.15
+  corridor: 0.2
+  persistence: 0.1
+  disturbance: 0.1
+  world_model: 0.25
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.2
+  planner_risk: 0.1
+  planner_ranking: 0.1
+  proposal_reconstruction: 0.2
+  proposal_success: 0.1
+  proposal_ranking: 0.1
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage1_dummy_seed15.yaml

@@ -0,0 +1,147 @@
+experiment_name: proxy_interaction_r3d_stage1_dummy_seed15
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 15
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 96
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_v6_rgbd_stage1_dummy_seed15.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_v6_rgbd_stage1_dummy_seed15.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 15
+optim:
+  epochs: 4
+  batch_size: 16
+  num_workers: 4
+  lr: 0.001
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: false
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 192
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: true
+  fusion:
+    hidden_dim: 192
+    num_cameras: 3
+    num_layers: 2
+    num_heads: 4
+    ff_dim: 384
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 192
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 4
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 192
+    num_heads: 4
+    num_layers: 2
+    ff_dim: 384
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 192
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 4
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 192
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 4
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 192
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 4
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.15
+  arm_role: 0.2
+  support_mode: 0.15
+  corridor: 0.2
+  persistence: 0.1
+  disturbance: 0.1
+  world_model: 0.25
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.2
+  planner_risk: 0.1
+  planner_ranking: 0.1
+  proposal_reconstruction: 0.2
+  proposal_success: 0.1
+  proposal_ranking: 0.1
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage2_clip_seed11.yaml

@@ -0,0 +1,149 @@
+experiment_name: proxy_interaction_r3d_stage2_clip_seed11
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 11
+init_checkpoint: /workspace/VLAarchtests/artifacts/outputs/reveal_runs/proxy_backbone_only_clip/checkpoint_best.pt
+init_strict: false
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 224
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_clip224_v6_rgbd_stage2_seed11.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_clip224_v6_rgbd_stage2_seed11.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 11
+optim:
+  epochs: 4
+  batch_size: 2
+  num_workers: 4
+  lr: 0.0003
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: true
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 512
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: false
+  fusion:
+    hidden_dim: 512
+    num_cameras: 3
+    num_layers: 4
+    num_heads: 8
+    ff_dim: 2048
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 512
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 8
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 512
+    num_heads: 8
+    num_layers: 4
+    ff_dim: 2048
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 512
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 8
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 512
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 512
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 8
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.1
+  arm_role: 0.15
+  support_mode: 0.1
+  corridor: 0.15
+  persistence: 0.05
+  disturbance: 0.05
+  world_model: 0.25
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.25
+  planner_risk: 0.1
+  planner_ranking: 0.2
+  proposal_reconstruction: 0.1
+  proposal_success: 0.15
+  proposal_ranking: 0.2
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage2_clip_seed12.yaml

@@ -0,0 +1,149 @@
+experiment_name: proxy_interaction_r3d_stage2_clip_seed12
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 12
+init_checkpoint: /workspace/VLAarchtests/artifacts/outputs/reveal_runs/proxy_backbone_only_clip/checkpoint_best.pt
+init_strict: false
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 224
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_clip224_v6_rgbd_stage2_seed12.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_clip224_v6_rgbd_stage2_seed12.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 12
+optim:
+  epochs: 4
+  batch_size: 2
+  num_workers: 4
+  lr: 0.0003
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: true
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 512
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: false
+  fusion:
+    hidden_dim: 512
+    num_cameras: 3
+    num_layers: 4
+    num_heads: 8
+    ff_dim: 2048
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 512
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 8
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 512
+    num_heads: 8
+    num_layers: 4
+    ff_dim: 2048
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 512
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 8
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 512
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 512
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 8
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.1
+  arm_role: 0.15
+  support_mode: 0.1
+  corridor: 0.15
+  persistence: 0.05
+  disturbance: 0.05
+  world_model: 0.25
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.25
+  planner_risk: 0.1
+  planner_ranking: 0.2
+  proposal_reconstruction: 0.1
+  proposal_success: 0.15
+  proposal_ranking: 0.2
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage2_clip_seed13.yaml

@@ -0,0 +1,149 @@
+experiment_name: proxy_interaction_r3d_stage2_clip_seed13
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 13
+init_checkpoint: /workspace/VLAarchtests/artifacts/outputs/reveal_runs/proxy_backbone_only_clip/checkpoint_best.pt
+init_strict: false
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 224
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_clip224_v6_rgbd_stage2_seed13.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_clip224_v6_rgbd_stage2_seed13.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 13
+optim:
+  epochs: 4
+  batch_size: 2
+  num_workers: 4
+  lr: 0.0003
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: true
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 512
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: false
+  fusion:
+    hidden_dim: 512
+    num_cameras: 3
+    num_layers: 4
+    num_heads: 8
+    ff_dim: 2048
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 512
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 8
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 512
+    num_heads: 8
+    num_layers: 4
+    ff_dim: 2048
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 512
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 8
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 512
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 512
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 8
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.1
+  arm_role: 0.15
+  support_mode: 0.1
+  corridor: 0.15
+  persistence: 0.05
+  disturbance: 0.05
+  world_model: 0.25
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.25
+  planner_risk: 0.1
+  planner_ranking: 0.2
+  proposal_reconstruction: 0.1
+  proposal_success: 0.15
+  proposal_ranking: 0.2
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage2_dummy_seed21.yaml

@@ -0,0 +1,148 @@
+experiment_name: proxy_interaction_r3d_stage2_dummy_seed21
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 21
+defaults: []
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 96
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_v6_rgbd_stage2_dummy_seed21.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_v6_rgbd_stage2_dummy_seed21.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 21
+optim:
+  epochs: 4
+  batch_size: 16
+  num_workers: 4
+  lr: 0.001
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: false
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 192
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: true
+  fusion:
+    hidden_dim: 192
+    num_cameras: 3
+    num_layers: 2
+    num_heads: 4
+    ff_dim: 384
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 192
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 4
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 192
+    num_heads: 4
+    num_layers: 2
+    ff_dim: 384
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 192
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 4
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 192
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 4
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 192
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 4
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.15
+  arm_role: 0.2
+  support_mode: 0.15
+  corridor: 0.2
+  persistence: 0.1
+  disturbance: 0.1
+  world_model: 0.3
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.2
+  planner_risk: 0.1
+  planner_ranking: 0.1
+  proposal_reconstruction: 0.2
+  proposal_success: 0.1
+  proposal_ranking: 0.1
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage2_dummy_seed22.yaml

@@ -0,0 +1,148 @@
+experiment_name: proxy_interaction_r3d_stage2_dummy_seed22
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 22
+defaults: []
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 96
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_v6_rgbd_stage2_dummy_seed22.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_v6_rgbd_stage2_dummy_seed22.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 22
+optim:
+  epochs: 4
+  batch_size: 16
+  num_workers: 4
+  lr: 0.001
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: false
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 192
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: true
+  fusion:
+    hidden_dim: 192
+    num_cameras: 3
+    num_layers: 2
+    num_heads: 4
+    ff_dim: 384
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 192
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 4
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 192
+    num_heads: 4
+    num_layers: 2
+    ff_dim: 384
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 192
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 4
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 192
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 4
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 192
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 4
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.15
+  arm_role: 0.2
+  support_mode: 0.15
+  corridor: 0.2
+  persistence: 0.1
+  disturbance: 0.1
+  world_model: 0.3
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.2
+  planner_risk: 0.1
+  planner_ranking: 0.1
+  proposal_reconstruction: 0.2
+  proposal_success: 0.1
+  proposal_ranking: 0.1
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage2_dummy_seed23.yaml

@@ -0,0 +1,148 @@
+experiment_name: proxy_interaction_r3d_stage2_dummy_seed23
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 23
+defaults: []
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 96
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_v6_rgbd_stage2_dummy_seed23.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_v6_rgbd_stage2_dummy_seed23.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 23
+optim:
+  epochs: 4
+  batch_size: 16
+  num_workers: 4
+  lr: 0.001
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: false
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: false
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 192
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: true
+  fusion:
+    hidden_dim: 192
+    num_cameras: 3
+    num_layers: 2
+    num_heads: 4
+    ff_dim: 384
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 192
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 4
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 192
+    num_heads: 4
+    num_layers: 2
+    ff_dim: 384
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 192
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 4
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 192
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 4
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 192
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 4
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.15
+  arm_role: 0.2
+  support_mode: 0.15
+  corridor: 0.2
+  persistence: 0.1
+  disturbance: 0.1
+  world_model: 0.3
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.2
+  planner_risk: 0.1
+  planner_ranking: 0.1
+  proposal_reconstruction: 0.2
+  proposal_success: 0.1
+  proposal_ranking: 0.1
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage3_clip_rgbd_seed17.yaml

@@ -0,0 +1,149 @@
+experiment_name: proxy_interaction_r3d_stage3_clip_rgbd_seed17
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 17
+init_checkpoint: /workspace/VLAarchtests/artifacts/outputs/reveal_runs/proxy_backbone_only_clip/checkpoint_best.pt
+init_strict: false
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 224
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_clip224_v6_rgbd_stage3_seed17.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_clip224_v6_rgbd_stage3_seed17.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 17
+optim:
+  epochs: 4
+  batch_size: 2
+  num_workers: 4
+  lr: 0.0003
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: true
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: true
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 512
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: false
+  fusion:
+    hidden_dim: 512
+    num_cameras: 3
+    num_layers: 4
+    num_heads: 8
+    ff_dim: 2048
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 512
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 8
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 512
+    num_heads: 8
+    num_layers: 4
+    ff_dim: 2048
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 512
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 8
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 512
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 512
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 8
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.1
+  arm_role: 0.15
+  support_mode: 0.1
+  corridor: 0.15
+  persistence: 0.05
+  disturbance: 0.05
+  world_model: 0.25
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.25
+  planner_risk: 0.1
+  planner_ranking: 0.2
+  proposal_reconstruction: 0.1
+  proposal_success: 0.15
+  proposal_ranking: 0.2
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage3_clip_rgbd_seed18.yaml

@@ -0,0 +1,149 @@
+experiment_name: proxy_interaction_r3d_stage3_clip_rgbd_seed18
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 18
+init_checkpoint: /workspace/VLAarchtests/artifacts/outputs/reveal_runs/proxy_backbone_only_clip/checkpoint_best.pt
+init_strict: false
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 224
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_clip224_v6_rgbd_stage3_seed18.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_clip224_v6_rgbd_stage3_seed18.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 18
+optim:
+  epochs: 4
+  batch_size: 2
+  num_workers: 4
+  lr: 0.0003
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: true
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: true
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 512
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: false
+  fusion:
+    hidden_dim: 512
+    num_cameras: 3
+    num_layers: 4
+    num_heads: 8
+    ff_dim: 2048
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 512
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 8
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 512
+    num_heads: 8
+    num_layers: 4
+    ff_dim: 2048
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 512
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 8
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 512
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 512
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 8
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.1
+  arm_role: 0.15
+  support_mode: 0.1
+  corridor: 0.15
+  persistence: 0.05
+  disturbance: 0.05
+  world_model: 0.25
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.25
+  planner_risk: 0.1
+  planner_ranking: 0.2
+  proposal_reconstruction: 0.1
+  proposal_success: 0.15
+  proposal_ranking: 0.2
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/generated_configs/proxy_interaction_r3d_stage3_clip_rgbd_seed19.yaml

@@ -0,0 +1,149 @@
+experiment_name: proxy_interaction_r3d_stage3_clip_rgbd_seed19
+output_dir: /workspace/VLAarchtests/artifacts/outputs/r3d
+device: cuda
+seed: 19
+init_checkpoint: /workspace/VLAarchtests/artifacts/outputs/reveal_runs/proxy_backbone_only_clip/checkpoint_best.pt
+init_strict: false
+data:
+  proxies:
+  - foliage_proxy
+  - bag_proxy
+  - cloth_proxy
+  resolution: 224
+  dataset_version: reveal_proxy_v6_rgbd_elastic_state
+  train_episodes_per_proxy: 48
+  val_episodes_per_proxy: 16
+  train_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_train_clip224_v6_rgbd_stage3_seed19.pt
+  val_dataset_path: /workspace/VLAarchtests/artifacts/data/reveal_proxy/proxy_val_clip224_v6_rgbd_stage3_seed19.pt
+  rebuild_dataset: false
+  chunk_horizon: 8
+  rollout_horizon: 5
+  history_steps: 6
+  planner_candidates: 8
+  seed: 19
+optim:
+  epochs: 4
+  batch_size: 2
+  num_workers: 4
+  lr: 0.0003
+  weight_decay: 0.0001
+trainer:
+  policy_type: elastic_reveal
+  use_bf16: true
+  grad_clip_norm: 1.0
+  freeze_backbone: true
+  gradient_checkpointing: false
+  plan_during_train: true
+  plan_during_eval: true
+  support_mode_conditioning: true
+  planner_mode: trainable
+  use_depth: true
+  use_world_model: true
+  use_role_tokens: true
+  compute_equivariance_probe: true
+policy:
+  backbone:
+    model_name: openai/clip-vit-base-patch32
+    hidden_dim: 512
+    max_text_tokens: 32
+    freeze_backbone: true
+    gradient_checkpointing: false
+    use_dummy_backbone: false
+  fusion:
+    hidden_dim: 512
+    num_cameras: 3
+    num_layers: 4
+    num_heads: 8
+    ff_dim: 2048
+    dropout: 0.1
+    proprio_dim: 32
+    proprio_tokens: 1
+  memory:
+    hidden_dim: 512
+    action_dim: 14
+    history_steps: 6
+    scene_history_steps: 3
+    belief_history_steps: 8
+    num_layers: 2
+    dropout: 0.1
+    memory_bank_size: 4
+    scene_bank_size: 2
+    belief_bank_size: 2
+    num_heads: 8
+    max_history_steps: 8
+  decoder:
+    hidden_dim: 512
+    num_heads: 8
+    num_layers: 4
+    ff_dim: 2048
+    dropout: 0.1
+    chunk_size: 8
+    action_dim: 14
+    arm_action_dim: 7
+    num_candidates: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_proposal_modes: 6
+    planner_top_k: 4
+  reveal_head:
+    hidden_dim: 512
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    belief_map_size: 32
+    field_size: 16
+    num_heads: 8
+    predict_belief_map: true
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+  world_model:
+    hidden_dim: 512
+    action_dim: 14
+    num_support_modes: 3
+    num_approach_templates: 32
+    rollout_horizon: 5
+    field_size: 16
+    num_heads: 8
+    num_phases: 5
+    num_arm_roles: 4
+    num_interaction_tokens: 8
+    belief_map_size: 32
+    predict_belief_map: true
+    scene_bank_size: 2
+    belief_bank_size: 2
+  planner:
+    hidden_dim: 512
+    num_candidates: 8
+    action_dim: 14
+    num_support_modes: 3
+    utility_margin: 0.1
+    num_heads: 8
+    num_layers: 2
+    num_phases: 5
+    num_arm_roles: 4
+    top_k: 4
+loss_weights:
+  action: 1.0
+  phase: 0.1
+  arm_role: 0.15
+  support_mode: 0.1
+  corridor: 0.15
+  persistence: 0.05
+  disturbance: 0.05
+  world_model: 0.25
+  belief: 0.05
+  visibility: 0.05
+  clearance: 0.05
+  support_stability: 0.05
+  reocclusion: 0.05
+  occluder_contact: 0.05
+  grasp_affordance: 0.05
+  planner_success: 0.25
+  planner_risk: 0.1
+  planner_ranking: 0.2
+  proposal_reconstruction: 0.1
+  proposal_success: 0.15
+  proposal_ranking: 0.2
+  proposal_diversity: 0.05
+  role_swap_consistency: 0.05

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_dual_push_buttons/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_dual_push_buttons --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_dual_push_buttons/stderr.txt

@@ -0,0 +1,2 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_dual_push_buttons/stdout.txt

@@ -0,0 +1,38 @@
+{
+  "display": ":99",
+  "headless": true,
+  "task": "BimanualDualPushButtons",
+  "description": "push the olive and the orange buttons",
+  "front_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_left_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_right_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "right_pose_shape": [
+    7
+  ],
+  "left_pose_shape": [
+    7
+  ],
+  "stepped_mode": "bimanual_noop",
+  "action_finite": true,
+  "action_dim": 18,
+  "reward": 0.0,
+  "done": false,
+  "front_rgb_shape_after_step": [
+    224,
+    224,
+    3
+  ]
+}
+[CoppeliaSim:loadinfo]   done.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_handover_item/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_handover_item --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_handover_item/stderr.txt

@@ -0,0 +1,2 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_handover_item/stdout.txt

@@ -0,0 +1,38 @@
+{
+  "display": ":99",
+  "headless": true,
+  "task": "BimanualHandoverItem",
+  "description": "bring me the red item",
+  "front_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_left_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_right_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "right_pose_shape": [
+    7
+  ],
+  "left_pose_shape": [
+    7
+  ],
+  "stepped_mode": "bimanual_noop",
+  "action_finite": true,
+  "action_dim": 18,
+  "reward": 0.0,
+  "done": false,
+  "front_rgb_shape_after_step": [
+    224,
+    224,
+    3
+  ]
+}
+[CoppeliaSim:loadinfo]   done.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_handover_item_easy/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_handover_item_easy --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_handover_item_easy/stderr.txt

@@ -0,0 +1,2 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_handover_item_easy/stdout.txt

@@ -0,0 +1,38 @@
+{
+  "display": ":99",
+  "headless": true,
+  "task": "BimanualHandoverItemEasy",
+  "description": "bring me the item",
+  "front_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_left_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_right_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "right_pose_shape": [
+    7
+  ],
+  "left_pose_shape": [
+    7
+  ],
+  "stepped_mode": "bimanual_noop",
+  "action_finite": true,
+  "action_dim": 18,
+  "reward": 0.0,
+  "done": false,
+  "front_rgb_shape_after_step": [
+    224,
+    224,
+    3
+  ]
+}
+[CoppeliaSim:loadinfo]   done.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_lift_ball/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_lift_ball --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_lift_ball/stderr.txt

@@ -0,0 +1,2 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_lift_ball/stdout.txt

@@ -0,0 +1,38 @@
+{
+  "display": ":99",
+  "headless": true,
+  "task": "BimanualLiftBall",
+  "description": "Lift the ball",
+  "front_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_left_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_right_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "right_pose_shape": [
+    7
+  ],
+  "left_pose_shape": [
+    7
+  ],
+  "stepped_mode": "bimanual_noop",
+  "action_finite": true,
+  "action_dim": 18,
+  "reward": 0.0,
+  "done": false,
+  "front_rgb_shape_after_step": [
+    224,
+    224,
+    3
+  ]
+}
+[CoppeliaSim:loadinfo]   done.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_lift_tray/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_lift_tray --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_lift_tray/stderr.txt

@@ -0,0 +1,2 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_lift_tray/stdout.txt

@@ -0,0 +1,38 @@
+{
+  "display": ":99",
+  "headless": true,
+  "task": "BimanualLiftTray",
+  "description": "Lift the tray",
+  "front_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_left_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_right_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "right_pose_shape": [
+    7
+  ],
+  "left_pose_shape": [
+    7
+  ],
+  "stepped_mode": "bimanual_noop",
+  "action_finite": true,
+  "action_dim": 18,
+  "reward": 0.0,
+  "done": false,
+  "front_rgb_shape_after_step": [
+    224,
+    224,
+    3
+  ]
+}
+[CoppeliaSim:loadinfo]   done.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_pick_laptop/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_pick_laptop --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_pick_laptop/stderr.txt

@@ -0,0 +1,6 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.
+WARNING:root:Waypoints are not reachable right=(False, 'waypoint1 - DualPanda') left=(True, '')
+ERROR:root:Error when checking waypoints. Exception is: Error in task bimanual_pick_laptop. Infeasible episode. Can't reach waypoint ((False, 'waypoint1 - DualPanda'), (True, '')).
+WARNING:root:Waypoints are not reachable right=(False, 'waypoint1 - DualPanda') left=(True, '')
+ERROR:root:Error when checking waypoints. Exception is: Error in task bimanual_pick_laptop. Infeasible episode. Can't reach waypoint ((False, 'waypoint1 - DualPanda'), (True, '')).

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_pick_laptop/stdout.txt

@@ -0,0 +1,38 @@
+{
+  "display": ":99",
+  "headless": true,
+  "task": "BimanualPickLaptop",
+  "description": "pick up the laptop",
+  "front_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_left_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_right_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "right_pose_shape": [
+    7
+  ],
+  "left_pose_shape": [
+    7
+  ],
+  "stepped_mode": "bimanual_noop",
+  "action_finite": true,
+  "action_dim": 18,
+  "reward": 0.0,
+  "done": false,
+  "front_rgb_shape_after_step": [
+    224,
+    224,
+    3
+  ]
+}
+[CoppeliaSim:loadinfo]   done.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_pick_plate/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_pick_plate --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_pick_plate/stderr.txt

@@ -0,0 +1,2 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_pick_plate/stdout.txt

@@ -0,0 +1,38 @@
+{
+  "display": ":99",
+  "headless": true,
+  "task": "BimanualPickPlate",
+  "description": "pick up the plate",
+  "front_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_left_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_right_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "right_pose_shape": [
+    7
+  ],
+  "left_pose_shape": [
+    7
+  ],
+  "stepped_mode": "bimanual_noop",
+  "action_finite": true,
+  "action_dim": 18,
+  "reward": 0.0,
+  "done": false,
+  "front_rgb_shape_after_step": [
+    224,
+    224,
+    3
+  ]
+}
+[CoppeliaSim:loadinfo]   done.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_push_box/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_push_box --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_push_box/stderr.txt

@@ -0,0 +1,2 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_push_box/stdout.txt

@@ -0,0 +1,38 @@
+{
+  "display": ":99",
+  "headless": true,
+  "task": "BimanualPushBox",
+  "description": "push the box to the red area",
+  "front_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_left_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_right_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "right_pose_shape": [
+    7
+  ],
+  "left_pose_shape": [
+    7
+  ],
+  "stepped_mode": "bimanual_noop",
+  "action_finite": true,
+  "action_dim": 18,
+  "reward": 0.0,
+  "done": false,
+  "front_rgb_shape_after_step": [
+    224,
+    224,
+    3
+  ]
+}
+[CoppeliaSim:loadinfo]   done.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_put_bottle_in_fridge/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_put_bottle_in_fridge --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_put_bottle_in_fridge/stderr.txt

@@ -0,0 +1,2 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_put_bottle_in_fridge/stdout.txt

@@ -0,0 +1,39 @@
+[ 0.05       -0.27500001  0.75199997]
+{
+  "display": ":99",
+  "headless": true,
+  "task": "BimanualPutBottleInFridge",
+  "description": "put bottle in fridge",
+  "front_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_left_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_right_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "right_pose_shape": [
+    7
+  ],
+  "left_pose_shape": [
+    7
+  ],
+  "stepped_mode": "bimanual_noop",
+  "action_finite": true,
+  "action_dim": 18,
+  "reward": 0.0,
+  "done": false,
+  "front_rgb_shape_after_step": [
+    224,
+    224,
+    3
+  ]
+}
+[CoppeliaSim:loadinfo]   done.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_put_item_in_drawer/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_put_item_in_drawer --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_put_item_in_drawer/stderr.txt

@@ -0,0 +1,12 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.
+WARNING:root:Waypoints are not reachable right=(False, 'waypoint4 - DualPanda') left=(False, 'waypoint2 - DualPanda')
+ERROR:root:Error when checking waypoints. Exception is: Error in task bimanual_put_item_in_drawer. Infeasible episode. Can't reach waypoint ((False, 'waypoint4 - DualPanda'), (False, 'waypoint2 - DualPanda')).
+WARNING:root:Waypoints are not reachable right=(False, 'waypoint4 - DualPanda') left=(False, 'waypoint2 - DualPanda')
+ERROR:root:Error when checking waypoints. Exception is: Error in task bimanual_put_item_in_drawer. Infeasible episode. Can't reach waypoint ((False, 'waypoint4 - DualPanda'), (False, 'waypoint2 - DualPanda')).
+WARNING:root:Waypoints are not reachable right=(False, 'waypoint4 - DualPanda') left=(True, '')
+ERROR:root:Error when checking waypoints. Exception is: Error in task bimanual_put_item_in_drawer. Infeasible episode. Can't reach waypoint ((False, 'waypoint4 - DualPanda'), (True, '')).
+WARNING:root:Waypoints are not reachable right=(False, 'waypoint4 - DualPanda') left=(False, 'waypoint2 - DualPanda')
+ERROR:root:Error when checking waypoints. Exception is: Error in task bimanual_put_item_in_drawer. Infeasible episode. Can't reach waypoint ((False, 'waypoint4 - DualPanda'), (False, 'waypoint2 - DualPanda')).
+WARNING:root:Waypoints are not reachable right=(False, 'waypoint4 - DualPanda') left=(True, '')
+ERROR:root:Error when checking waypoints. Exception is: Error in task bimanual_put_item_in_drawer. Infeasible episode. Can't reach waypoint ((False, 'waypoint4 - DualPanda'), (True, '')).

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_put_item_in_drawer/stdout.txt

@@ -0,0 +1,38 @@
+{
+  "display": ":99",
+  "headless": true,
+  "task": "BimanualPutItemInDrawer",
+  "description": "put the item in the bottom drawer",
+  "front_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_left_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_right_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "right_pose_shape": [
+    7
+  ],
+  "left_pose_shape": [
+    7
+  ],
+  "stepped_mode": "bimanual_noop",
+  "action_finite": true,
+  "action_dim": 18,
+  "reward": 0.0,
+  "done": false,
+  "front_rgb_shape_after_step": [
+    224,
+    224,
+    3
+  ]
+}
+[CoppeliaSim:loadinfo]   done.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_straighten_rope/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_straighten_rope --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_straighten_rope/stderr.txt

@@ -0,0 +1,4 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.
+WARNING:root:Waypoints are not reachable right=(False, 'waypoint3 - DualPanda') left=(True, '')
+ERROR:root:Error when checking waypoints. Exception is: Error in task bimanual_straighten_rope. Infeasible episode. Can't reach waypoint ((False, 'waypoint3 - DualPanda'), (True, '')).

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_straighten_rope/stdout.txt

@@ -0,0 +1,38 @@
+{
+  "display": ":99",
+  "headless": true,
+  "task": "BimanualStraightenRope",
+  "description": "straighten rope",
+  "front_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_left_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "wrist_right_rgb_shape": [
+    224,
+    224,
+    3
+  ],
+  "right_pose_shape": [
+    7
+  ],
+  "left_pose_shape": [
+    7
+  ],
+  "stepped_mode": "bimanual_noop",
+  "action_finite": true,
+  "action_dim": 18,
+  "reward": 0.0,
+  "done": false,
+  "front_rgb_shape_after_step": [
+    224,
+    224,
+    3
+  ]
+}
+[CoppeliaSim:loadinfo]   done.

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_sweep_to_dustpan/command.txt

@@ -0,0 +1 @@
+/workspace/envs/rlbench/bin/python -m sim_rlbench.launch_smoke --task bimanual_sweep_to_dustpan --resolution 224 --headless

artifacts/outputs/r3d/peract2_13_launch_smoke/bimanual_sweep_to_dustpan/stderr.txt

@@ -0,0 +1,2 @@
+qt.qpa.xcb: QXcbConnection: XCB error: 148 (Unknown), sequence: 181, resource id: 0, major code: 140 (Unknown), minor code: 20
+WARNING:root:not sure how _robot_shapes are used is used.