Import checkpoint from kimtaey/rldx-1-image

config.json

@@ -0,0 +1,105 @@
+{
+  "action_horizon": 16,
+  "action_model_lora_alpha": 32,
+  "action_model_lora_dropout": 0.0,
+  "action_model_lora_rank": 16,
+  "action_model_lora_target_modules": [
+    "vl_qkv",
+    "vl_proj",
+    "sa_qkv",
+    "sa_proj",
+    "p_qkv",
+    "p_proj",
+    "linear1",
+    "linear2"
+  ],
+  "action_model_use_lora": false,
+  "architectures": [
+    "RLDX"
+  ],
+  "backbone_lora_alpha": 32,
+  "backbone_lora_dropout": 0.0,
+  "backbone_lora_num_layers": -1,
+  "backbone_lora_rank": 16,
+  "backbone_lora_target_modules": [
+    "q_proj",
+    "k_proj",
+    "v_proj",
+    "o_proj",
+    "gate_proj",
+    "up_proj",
+    "down_proj"
+  ],
+  "backbone_use_lora": false,
+  "color_jitter_params": {
+    "brightness": 0.3,
+    "contrast": 0.4,
+    "hue": 0.08,
+    "saturation": 0.5
+  },
+  "conversation_image_first": false,
+  "diffusion_model_cfg": {
+    "action_model_max_seq_len": 512,
+    "attention_head_dim": 64,
+    "depth_multi_stream": 4,
+    "depth_single_stream": 8,
+    "dropout": 0.2,
+    "final_dropout": true,
+    "num_attention_heads": 24,
+    "output_dim": 1024,
+    "physics_dim": 0,
+    "positional_embeddings": "rope_sa_only",
+    "pre_norm": "layer_norm",
+    "qk_norm": "rms_norm",
+    "rope_theta": 10000.0,
+    "sa_dim": 1536,
+    "temb_type": "input_token",
+    "use_physics": false,
+    "use_swiglu": true,
+    "vl_dim": 4096
+  },
+  "dtype": "bfloat16",
+  "freeze_cog_tokens": false,
+  "general_embodiment_train_ratio": 0.02,
+  "image_max_area": 65536,
+  "image_resize_m": 32,
+  "memory_cfg": {
+    "hidden_size": 4096,
+    "intermediate_size": 16384,
+    "max_position_embeddings": 32,
+    "num_attention_heads": 16,
+    "num_hidden_layers": 2,
+    "num_key_value_heads": 16,
+    "rms_norm_eps": 1e-05,
+    "use_causal_attn": true,
+    "use_rope": true
+  },
+  "model_name": "/data/taeyoung/models/RLDX-1-VLM",
+  "model_revision": null,
+  "model_type": "RLDX-1",
+  "n_cog_tokens": 64,
+  "physics_delta_indices": null,
+  "physics_dims": [],
+  "physics_keys": [],
+  "physics_use_flow_matching": true,
+  "random_crop_fraction": null,
+  "random_rotation_angle": null,
+  "rtc_inference_delay": 0,
+  "rtc_inference_exec_horizon": 0,
+  "rtc_inference_mode": "none",
+  "rtc_jacobian_steps_only": 3,
+  "rtc_training_max_delay": 0,
+  "state_dropout_prob": 0.0,
+  "transformers_version": "4.57.0",
+  "tune_diffusion_model": true,
+  "tune_llm": false,
+  "tune_projector": true,
+  "tune_top_llm_layers": 4,
+  "tune_visual": false,
+  "use_memory": false,
+  "use_motion": false,
+  "use_physics": false,
+  "use_video": true,
+  "video_length": 1,
+  "video_stride": 2
+}

experiment_cfg/conf.yaml

@@ -0,0 +1,2661 @@
+load_config_path: null
+model:
+  model_type: RLDX-1
+  model_dtype: bfloat16
+  model_name: /data/taeyoung/models/RLDX-1-VLM
+  backbone_model_type: vtc_qwen3_vl
+  model_revision: null
+  tune_top_llm_layers: 4
+  backbone_embedding_dim: 4096
+  tune_llm: false
+  tune_visual: false
+  select_layer: 18
+  reproject_vision: false
+  use_flash_attention: true
+  load_bf16: true
+  backbone_trainable_params_fp32: true
+  freeze_cog_tokens: false
+  image_max_area: 65536
+  image_resize_m: 32
+  random_crop_fraction: null
+  random_rotation_angle: null
+  color_jitter_params:
+    brightness: 0.3
+    contrast: 0.4
+    saturation: 0.5
+    hue: 0.08
+  formalize_language: true
+  apply_sincos_state_encoding: false
+  use_relative_action: true
+  use_percentiles: true
+  conversation_image_first: false
+  max_state_dim: 64
+  max_action_dim: 64
+  action_horizon: 16
+  hidden_size: 1024
+  input_embedding_dim: 1536
+  general_embodiment_train_ratio: 0.02
+  add_pos_embed: true
+  attn_dropout: 0.2
+  use_vlln: false
+  max_seq_len: 1024
+  n_cog_tokens: 64
+  diffusion_model_cfg:
+    attention_head_dim: 64
+    depth_multi_stream: 4
+    depth_single_stream: 8
+    dropout: 0.2
+    final_dropout: true
+    num_attention_heads: 24
+    output_dim: 1024
+    positional_embeddings: rope_sa_only
+    rope_theta: 10000.0
+    temb_type: input_token
+    use_swiglu: true
+    action_model_max_seq_len: 512
+    pre_norm: layer_norm
+    qk_norm: rms_norm
+  num_inference_timesteps: 4
+  noise_beta_alpha: 1.5
+  noise_beta_beta: 1.0
+  noise_s: 0.999
+  num_timestep_buckets: 1000
+  tune_projector: true
+  tune_diffusion_model: true
+  tune_vlln: true
+  action_model_use_lora: false
+  action_model_lora_rank: 16
+  action_model_lora_alpha: 32
+  action_model_lora_dropout: 0.0
+  action_model_lora_target_modules:
+  - vl_qkv
+  - vl_proj
+  - sa_qkv
+  - sa_proj
+  - p_qkv
+  - p_proj
+  - linear1
+  - linear2
+  backbone_use_lora: false
+  backbone_lora_rank: 16
+  backbone_lora_alpha: 32
+  backbone_lora_dropout: 0.0
+  backbone_lora_num_layers: -1
+  backbone_lora_target_modules:
+  - q_proj
+  - k_proj
+  - v_proj
+  - o_proj
+  - gate_proj
+  - up_proj
+  - down_proj
+  state_dropout_prob: 0.0
+  state_additive_noise_scale: 0.0
+  rtc_training_max_delay: 0
+  rtc_inference_mode: none
+  rtc_inference_delay: 0
+  rtc_inference_exec_horizon: 0
+  rtc_jacobian_steps_only: 3
+  use_memory: false
+  memory_length: 4
+  memory_n_cog_tokens: null
+  concat_memory: false
+  memory_dropout_prob: 0.0
+  memory_stride: 16
+  memory_cfg:
+    hidden_size: 4096
+    intermediate_size: 16384
+    num_hidden_layers: 2
+    num_attention_heads: 16
+    num_key_value_heads: 16
+    max_position_embeddings: 32
+    rms_norm_eps: 1.0e-05
+    use_causal_attn: true
+    use_rope: true
+  use_motion: false
+  motion_insert_layer: 9
+  motion_d_hid: 512
+  motion_window:
+  - 5
+  - 9
+  - 9
+  motion_ext_chnls:
+  - 256
+  motion_int_chnls:
+  - 256
+  - 256
+  - 512
+  motion_corr_func: cosine
+  motion_n_encoders: 1
+  motion_use_layerscale: false
+  motion_layerscale_init: 1.0e-05
+  motion_use_layernorm: false
+  motion_use_syncbn: false
+  motion_injection_point: vision_encoder
+  motion_pool_type: avg
+  motion_drop: true
+  motion_gradient_check: false
+  motion_int_mode: lite
+  use_video: true
+  video_length: 1
+  video_stride: 2
+  use_physics: false
+  physics_keys: []
+  physics_dims: []
+  physics_loss_weight: 0.1
+  allow_missing_physics: false
+  physics_delta_indices: null
+  physics_use_flow_matching: true
+  physics_dropout_prob: 0.0
+  max_num_embodiments: 36
+data:
+  datasets:
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/fractal20220817_data
+    embodiment_tag: fractal20220817_data
+    mix_ratio: 0.04
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/kuka
+    embodiment_tag: kuka
+    mix_ratio: 0.04
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/bridge_orig
+    embodiment_tag: bridge_orig
+    mix_ratio: 0.04
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/taco_play
+    embodiment_tag: taco_play
+    mix_ratio: 0.0084
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/jaco_play
+    embodiment_tag: jaco_play
+    mix_ratio: 0.0015
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/berkeley_cable_routing
+    embodiment_tag: berkeley_cable_routing
+    mix_ratio: 0.0008
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/roboturk
+    embodiment_tag: roboturk
+    mix_ratio: 0.0065
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/viola
+    embodiment_tag: viola
+    mix_ratio: 0.003
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/berkeley_autolab_ur5
+    embodiment_tag: berkeley_autolab_ur5
+    mix_ratio: 0.003
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/toto
+    embodiment_tag: toto
+    mix_ratio: 0.005
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/language_table
+    embodiment_tag: language_table
+    mix_ratio: 0.0125
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/stanford_hydra_dataset_converted_externally_to_rlds
+    embodiment_tag: stanford_hydra_dataset_converted_externally_to_rlds
+    mix_ratio: 0.0125
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/austin_buds_dataset_converted_externally_to_rlds
+    embodiment_tag: austin_buds_dataset_converted_externally_to_rlds
+    mix_ratio: 0.0006
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/nyu_franka_play_dataset_converted_externally_to_rlds
+    embodiment_tag: nyu_franka_play_dataset_converted_externally_to_rlds
+    mix_ratio: 0.002
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/furniture_bench_dataset_converted_externally_to_rlds
+    embodiment_tag: furniture_bench_dataset_converted_externally_to_rlds
+    mix_ratio: 0.007
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/ucsd_kitchen_dataset_converted_externally_to_rlds
+    embodiment_tag: ucsd_kitchen_dataset_converted_externally_to_rlds
+    mix_ratio: 0.0001
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/austin_sailor_dataset_converted_externally_to_rlds
+    embodiment_tag: austin_sailor_dataset_converted_externally_to_rlds
+    mix_ratio: 0.006
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/austin_sirius_dataset_converted_externally_to_rlds
+    embodiment_tag: austin_sirius_dataset_converted_externally_to_rlds
+    mix_ratio: 0.005
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/dlr_edan_shared_control_converted_externally_to_rlds
+    embodiment_tag: dlr_edan_shared_control_converted_externally_to_rlds
+    mix_ratio: 0.0002
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/iamlab_cmu_pickup_insert_converted_externally_to_rlds
+    embodiment_tag: iamlab_cmu_pickup_insert_converted_externally_to_rlds
+    mix_ratio: 0.0025
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/utaustin_mutex
+    embodiment_tag: utaustin_mutex
+    mix_ratio: 0.006
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/berkeley_fanuc_manipulation
+    embodiment_tag: berkeley_fanuc_manipulation
+    mix_ratio: 0.002
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/cmu_stretch
+    embodiment_tag: cmu_stretch
+    mix_ratio: 0.0004
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/bc_z
+    embodiment_tag: bc_z
+    mix_ratio: 0.02
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/fmb_dataset
+    embodiment_tag: fmb_dataset
+    mix_ratio: 0.02
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/dobbe
+    embodiment_tag: dobbe
+    mix_ratio: 0.005
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/droid
+    embodiment_tag: droid
+    mix_ratio: 0.25
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/galaxea_part1
+    embodiment_tag: galaxea
+    mix_ratio: 0.02
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/galaxea_part2
+    embodiment_tag: galaxea
+    mix_ratio: 0.02
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/galaxea_part3
+    embodiment_tag: galaxea
+    mix_ratio: 0.02
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/galaxea_part4
+    embodiment_tag: galaxea
+    mix_ratio: 0.02
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/galaxea_part5
+    embodiment_tag: galaxea
+    mix_ratio: 0.02
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/agibot_gripper_part1
+    embodiment_tag: agibot_gripper
+    mix_ratio: 0.025
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/agibot_gripper_part2
+    embodiment_tag: agibot_gripper
+    mix_ratio: 0.025
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/agibot_gripper_part3
+    embodiment_tag: agibot_gripper
+    mix_ratio: 0.025
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/agibot_gripper_part4
+    embodiment_tag: agibot_gripper
+    mix_ratio: 0.025
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/agibot_dexhand
+    embodiment_tag: agibot_dexhand
+    mix_ratio: 0.05
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/action_net
+    embodiment_tag: action_net
+    mix_ratio: 0.1
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/march_robocurate_v1
+    embodiment_tag: neural_gr1
+    mix_ratio: 0.03
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/march_robocurate_v2
+    embodiment_tag: neural_gr1
+    mix_ratio: 0.025
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/march_robocurate_v3
+    embodiment_tag: neural_gr1
+    mix_ratio: 0.035
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/march_robocurate_v4
+    embodiment_tag: neural_gr1
+    mix_ratio: 0.01
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/humanoid_everyday_g1
+    embodiment_tag: humanoid_everyday_g1
+    mix_ratio: 0.025
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  - dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/humanoid_everyday_h1
+    embodiment_tag: humanoid_everyday_h1
+    mix_ratio: 0.025
+    dataset_type: physical_embodiment
+    val_dataset_path: null
+  modality_configs:
+    agibot_gripper:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist_left
+        - wrist_right
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - state
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - action
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    jaco_play:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    utaustin_mutex:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - joint_position
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    kuka:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    austin_sirius_dataset_converted_externally_to_rlds:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    bc_z:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    taco_play:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    galaxea:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist_left
+        - wrist_right
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - state
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - action
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    fmb_dataset:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - secondary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    agibot_dexhand:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - state
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - action
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    language_table:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    humanoid_everyday_g1:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - egocentric_resized
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - left_arm
+        - left_hand
+        - right_arm
+        - right_hand
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - left_arm
+        - left_hand
+        - right_arm
+        - right_hand
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    nyu_franka_play_dataset_converted_externally_to_rlds:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - secondary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    austin_buds_dataset_converted_externally_to_rlds:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - joint_position
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    dlr_edan_shared_control_converted_externally_to_rlds:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    droid:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - secondary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    berkeley_fanuc_manipulation:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - joint_position
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    dobbe:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    cmu_stretch:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    austin_sailor_dataset_converted_externally_to_rlds:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    furniture_bench_dataset_converted_externally_to_rlds:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    bridge_orig:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - secondary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    ucsd_kitchen_dataset_converted_externally_to_rlds:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - joint_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    toto:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - joint_position
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    neural_gr1:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - ego_view
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - left_arm
+        - left_hand
+        - left_leg
+        - neck
+        - right_arm
+        - right_hand
+        - right_leg
+        - waist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - left_arm
+        - left_hand
+        - left_leg
+        - neck
+        - right_arm
+        - right_hand
+        - right_leg
+        - waist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    berkeley_cable_routing:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - secondary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - joint_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    berkeley_autolab_ur5:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    iamlab_cmu_pickup_insert_converted_externally_to_rlds:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    stanford_hydra_dataset_converted_externally_to_rlds:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    humanoid_everyday_h1:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - egocentric_resized
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - left_arm
+        - left_hand
+        - right_arm
+        - right_hand
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - left_arm
+        - left_hand
+        - right_arm
+        - right_hand
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    fractal20220817_data:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    roboturk:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - none
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    action_net:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - state
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - action
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+    viola:
+      video:
+        delta_indices:
+        - 0
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      state:
+        delta_indices:
+        - 0
+        modality_keys:
+        - joint_position
+        - gripper_position
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+      action:
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs:
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: DELTA
+          type: EEF
+          format: DEFAULT
+          state_key: null
+        - rep: ABSOLUTE
+          type: NON_EEF
+          format: DEFAULT
+          state_key: null
+      language:
+        delta_indices:
+        - 0
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+        mean_std_embedding_keys: null
+        action_configs: null
+  dataset_mode: sharded
+  download_cache: false
+  shard_size: 1024
+  episode_sampling_rate: 0.1
+  num_shards_per_epoch: 100000
+  override_pretraining_statistics: false
+  mode: single_turn
+  random_chop: 0.0
+  mock_dataset_mode: false
+  shuffle: true
+  seed: 42
+  multiprocessing_context: fork
+  allow_padding: true
+  subsample_ratio: 1.0
+  video_backend: torchcodec
+training:
+  output_dir: /data/taeyoung/workspace/RLDX-1/ckpt/rldx1/pretrained/rldx1_image_pt_mixv5_bsz8192_100k
+  experiment_name: rldx1_image_pt_mixv5_bsz8192_100k
+  max_steps: 100000
+  global_batch_size: 4096
+  batch_size: null
+  gradient_accumulation_steps: 2
+  learning_rate: 0.0001
+  lr_scheduler_type: constant_with_warmup
+  weight_decay: 1.0e-05
+  warmup_ratio: 0.05
+  warmup_steps: 0
+  max_grad_norm: 1.0
+  new_param_warmup_steps: 0
+  optim: adamw_torch_fused
+  start_from_checkpoint: null
+  tf32: true
+  fp16: false
+  bf16: true
+  eval_bf16: true
+  logging_steps: 10
+  save_steps: 1000
+  save_total_limit: 40
+  save_vl_model: false
+  upload_checkpoints: false
+  upload_every: 1000
+  upload_last_n_checkpoints: 5
+  max_concurrent_uploads: 2
+  eval_strategy: 'no'
+  eval_steps: 500
+  eval_set_split_ratio: 0.1
+  eval_batch_size: 2
+  save_best_eval_metric_name: ''
+  save_best_eval_metric_greater_is_better: true
+  deepspeed_stage: 2
+  gradient_checkpointing: false
+  transformers_trust_remote_code: true
+  transformers_local_files_only: false
+  transformers_cache_dir: null
+  transformers_access_token: null
+  use_ddp: false
+  ddp_bucket_cap_mb: 100
+  num_gpus: 64
+  dataloader_num_workers: 40
+  remove_unused_columns: false
+  use_wandb: true
+  wandb_project: rldx-pretrain
+  enable_profiling: false
+  max_retries: 3
+  assert_loss_less_than: null
+  add_rl_callback: false
+  enable_open_loop_eval: false
+  open_loop_eval_traj_ids:
+  - 0
+  open_loop_eval_steps_per_traj: 100
+  open_loop_eval_plot_indices: null
+max_steps: 100000
+save_steps: 1000

experiment_cfg/config.yaml

@@ -0,0 +1,2808 @@
+!!python/object:rldx.configs.base_config.Config
+data: !!python/object:rldx.configs.data.data_config.DataConfig
+  allow_padding: true
+  dataset_mode: sharded
+  datasets:
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/fractal20220817_data
+    dataset_type: physical_embodiment
+    embodiment_tag: fractal20220817_data
+    mix_ratio: 0.04
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/kuka
+    dataset_type: physical_embodiment
+    embodiment_tag: kuka
+    mix_ratio: 0.04
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/bridge_orig
+    dataset_type: physical_embodiment
+    embodiment_tag: bridge_orig
+    mix_ratio: 0.04
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/taco_play
+    dataset_type: physical_embodiment
+    embodiment_tag: taco_play
+    mix_ratio: 0.0084
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/jaco_play
+    dataset_type: physical_embodiment
+    embodiment_tag: jaco_play
+    mix_ratio: 0.0015
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/berkeley_cable_routing
+    dataset_type: physical_embodiment
+    embodiment_tag: berkeley_cable_routing
+    mix_ratio: 0.0008
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/roboturk
+    dataset_type: physical_embodiment
+    embodiment_tag: roboturk
+    mix_ratio: 0.0065
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/viola
+    dataset_type: physical_embodiment
+    embodiment_tag: viola
+    mix_ratio: 0.003
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/berkeley_autolab_ur5
+    dataset_type: physical_embodiment
+    embodiment_tag: berkeley_autolab_ur5
+    mix_ratio: 0.003
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/toto
+    dataset_type: physical_embodiment
+    embodiment_tag: toto
+    mix_ratio: 0.005
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/language_table
+    dataset_type: physical_embodiment
+    embodiment_tag: language_table
+    mix_ratio: 0.0125
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/stanford_hydra_dataset_converted_externally_to_rlds
+    dataset_type: physical_embodiment
+    embodiment_tag: stanford_hydra_dataset_converted_externally_to_rlds
+    mix_ratio: 0.0125
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/austin_buds_dataset_converted_externally_to_rlds
+    dataset_type: physical_embodiment
+    embodiment_tag: austin_buds_dataset_converted_externally_to_rlds
+    mix_ratio: 0.0006
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/nyu_franka_play_dataset_converted_externally_to_rlds
+    dataset_type: physical_embodiment
+    embodiment_tag: nyu_franka_play_dataset_converted_externally_to_rlds
+    mix_ratio: 0.002
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/furniture_bench_dataset_converted_externally_to_rlds
+    dataset_type: physical_embodiment
+    embodiment_tag: furniture_bench_dataset_converted_externally_to_rlds
+    mix_ratio: 0.007
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/ucsd_kitchen_dataset_converted_externally_to_rlds
+    dataset_type: physical_embodiment
+    embodiment_tag: ucsd_kitchen_dataset_converted_externally_to_rlds
+    mix_ratio: 0.0001
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/austin_sailor_dataset_converted_externally_to_rlds
+    dataset_type: physical_embodiment
+    embodiment_tag: austin_sailor_dataset_converted_externally_to_rlds
+    mix_ratio: 0.006
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/austin_sirius_dataset_converted_externally_to_rlds
+    dataset_type: physical_embodiment
+    embodiment_tag: austin_sirius_dataset_converted_externally_to_rlds
+    mix_ratio: 0.005
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/dlr_edan_shared_control_converted_externally_to_rlds
+    dataset_type: physical_embodiment
+    embodiment_tag: dlr_edan_shared_control_converted_externally_to_rlds
+    mix_ratio: 0.0002
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/iamlab_cmu_pickup_insert_converted_externally_to_rlds
+    dataset_type: physical_embodiment
+    embodiment_tag: iamlab_cmu_pickup_insert_converted_externally_to_rlds
+    mix_ratio: 0.0025
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/utaustin_mutex
+    dataset_type: physical_embodiment
+    embodiment_tag: utaustin_mutex
+    mix_ratio: 0.006
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/berkeley_fanuc_manipulation
+    dataset_type: physical_embodiment
+    embodiment_tag: berkeley_fanuc_manipulation
+    mix_ratio: 0.002
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/cmu_stretch
+    dataset_type: physical_embodiment
+    embodiment_tag: cmu_stretch
+    mix_ratio: 0.0004
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/bc_z
+    dataset_type: physical_embodiment
+    embodiment_tag: bc_z
+    mix_ratio: 0.02
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/fmb_dataset
+    dataset_type: physical_embodiment
+    embodiment_tag: fmb_dataset
+    mix_ratio: 0.02
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/dobbe
+    dataset_type: physical_embodiment
+    embodiment_tag: dobbe
+    mix_ratio: 0.005
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/droid
+    dataset_type: physical_embodiment
+    embodiment_tag: droid
+    mix_ratio: 0.25
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/galaxea_part1
+    dataset_type: physical_embodiment
+    embodiment_tag: galaxea
+    mix_ratio: 0.02
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/galaxea_part2
+    dataset_type: physical_embodiment
+    embodiment_tag: galaxea
+    mix_ratio: 0.02
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/galaxea_part3
+    dataset_type: physical_embodiment
+    embodiment_tag: galaxea
+    mix_ratio: 0.02
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/galaxea_part4
+    dataset_type: physical_embodiment
+    embodiment_tag: galaxea
+    mix_ratio: 0.02
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/galaxea_part5
+    dataset_type: physical_embodiment
+    embodiment_tag: galaxea
+    mix_ratio: 0.02
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/agibot_gripper_part1
+    dataset_type: physical_embodiment
+    embodiment_tag: agibot_gripper
+    mix_ratio: 0.025
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/agibot_gripper_part2
+    dataset_type: physical_embodiment
+    embodiment_tag: agibot_gripper
+    mix_ratio: 0.025
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/agibot_gripper_part3
+    dataset_type: physical_embodiment
+    embodiment_tag: agibot_gripper
+    mix_ratio: 0.025
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/agibot_gripper_part4
+    dataset_type: physical_embodiment
+    embodiment_tag: agibot_gripper
+    mix_ratio: 0.025
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/agibot_dexhand
+    dataset_type: physical_embodiment
+    embodiment_tag: agibot_dexhand
+    mix_ratio: 0.05
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/action_net
+    dataset_type: physical_embodiment
+    embodiment_tag: action_net
+    mix_ratio: 0.1
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/march_robocurate_v1
+    dataset_type: physical_embodiment
+    embodiment_tag: neural_gr1
+    mix_ratio: 0.03
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/march_robocurate_v2
+    dataset_type: physical_embodiment
+    embodiment_tag: neural_gr1
+    mix_ratio: 0.025
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/march_robocurate_v3
+    dataset_type: physical_embodiment
+    embodiment_tag: neural_gr1
+    mix_ratio: 0.035
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/march_robocurate_v4
+    dataset_type: physical_embodiment
+    embodiment_tag: neural_gr1
+    mix_ratio: 0.01
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/humanoid_everyday_g1
+    dataset_type: physical_embodiment
+    embodiment_tag: humanoid_everyday_g1
+    mix_ratio: 0.025
+    val_dataset_path: null
+  - !!python/object:rldx.configs.data.data_config.SingleDatasetConfig
+    dataset_paths:
+    - /data/taeyoung/data/vla_pretrain_dataset/openx_lerobot/humanoid_everyday_h1
+    dataset_type: physical_embodiment
+    embodiment_tag: humanoid_everyday_h1
+    mix_ratio: 0.025
+    val_dataset_path: null
+  download_cache: false
+  episode_sampling_rate: 0.1
+  mock_dataset_mode: false
+  modality_configs:
+    action_net:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: &id001 !!python/object/apply:rldx.data.types.ActionFormat
+          - default
+          rep: &id002 !!python/object/apply:rldx.data.types.ActionRepresentation
+          - absolute
+          state_key: null
+          type: &id003 !!python/object/apply:rldx.data.types.ActionType
+          - non_eef
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - action
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - state
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+    agibot_dexhand:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - action
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - state
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+    agibot_gripper:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - action
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - state
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist_left
+        - wrist_right
+        sin_cos_embedding_keys: null
+    austin_buds_dataset_converted_externally_to_rlds:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: &id004 !!python/object/apply:rldx.data.types.ActionRepresentation
+          - delta
+          state_key: null
+          type: &id005 !!python/object/apply:rldx.data.types.ActionType
+          - eef
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - joint_position
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+    austin_sailor_dataset_converted_externally_to_rlds:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+    austin_sirius_dataset_converted_externally_to_rlds:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+    bc_z:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+    berkeley_autolab_ur5:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+    berkeley_cable_routing:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - joint_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - secondary
+        - wrist
+        sin_cos_embedding_keys: null
+    berkeley_fanuc_manipulation:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - joint_position
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+    bridge_orig:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - secondary
+        sin_cos_embedding_keys: null
+    cmu_stretch:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+    dlr_edan_shared_control_converted_externally_to_rlds:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+    dobbe:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - wrist
+        sin_cos_embedding_keys: null
+    droid:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - secondary
+        - wrist
+        sin_cos_embedding_keys: null
+    fmb_dataset:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - secondary
+        - wrist
+        sin_cos_embedding_keys: null
+    fractal20220817_data:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+    furniture_bench_dataset_converted_externally_to_rlds:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+    galaxea:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - action
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - state
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist_left
+        - wrist_right
+        sin_cos_embedding_keys: null
+    humanoid_everyday_g1:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - left_arm
+        - left_hand
+        - right_arm
+        - right_hand
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - left_arm
+        - left_hand
+        - right_arm
+        - right_hand
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - egocentric_resized
+        sin_cos_embedding_keys: null
+    humanoid_everyday_h1:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - left_arm
+        - left_hand
+        - right_arm
+        - right_hand
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - left_arm
+        - left_hand
+        - right_arm
+        - right_hand
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - egocentric_resized
+        sin_cos_embedding_keys: null
+    iamlab_cmu_pickup_insert_converted_externally_to_rlds:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+    jaco_play:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+    kuka:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+    language_table:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+    neural_gr1:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - left_arm
+        - left_hand
+        - left_leg
+        - neck
+        - right_arm
+        - right_hand
+        - right_leg
+        - waist
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - left_arm
+        - left_hand
+        - left_leg
+        - neck
+        - right_arm
+        - right_hand
+        - right_leg
+        - waist
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - ego_view
+        sin_cos_embedding_keys: null
+    nyu_franka_play_dataset_converted_externally_to_rlds:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - secondary
+        sin_cos_embedding_keys: null
+    roboturk:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - none
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+    stanford_hydra_dataset_converted_externally_to_rlds:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+    taco_play:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+    toto:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - joint_position
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+    ucsd_kitchen_dataset_converted_externally_to_rlds:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - joint_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        sin_cos_embedding_keys: null
+    utaustin_mutex:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - joint_position
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+    viola:
+      action: !!python/object:rldx.data.types.ModalityConfig
+        action_configs:
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id004
+          state_key: null
+          type: *id005
+        - !!python/object:rldx.data.types.ActionConfig
+          format: *id001
+          rep: *id002
+          state_key: null
+          type: *id003
+        delta_indices:
+        - 0
+        - 1
+        - 2
+        - 3
+        - 4
+        - 5
+        - 6
+        - 7
+        - 8
+        - 9
+        - 10
+        - 11
+        - 12
+        - 13
+        - 14
+        - 15
+        mean_std_embedding_keys: null
+        modality_keys:
+        - end_effector_position
+        - end_effector_rotation
+        - gripper_close
+        sin_cos_embedding_keys: null
+      language: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - annotation.human.action.task_description
+        sin_cos_embedding_keys: null
+      state: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - joint_position
+        - gripper_position
+        sin_cos_embedding_keys: null
+      video: !!python/object:rldx.data.types.ModalityConfig
+        action_configs: null
+        delta_indices:
+        - 0
+        mean_std_embedding_keys: null
+        modality_keys:
+        - primary
+        - wrist
+        sin_cos_embedding_keys: null
+  mode: single_turn
+  multiprocessing_context: fork
+  num_shards_per_epoch: 100000
+  override_pretraining_statistics: false
+  random_chop: 0.0
+  seed: 42
+  shard_size: 1024
+  shuffle: true
+  subsample_ratio: 1.0
+  video_backend: torchcodec
+load_config_path: null
+model: !!python/object:rldx.configs.model.rldx.RLDXConfig
+  _attn_implementation_internal: null
+  _commit_hash: null
+  _name_or_path: ''
+  _output_attentions: false
+  action_horizon: 16
+  action_model_lora_alpha: 32
+  action_model_lora_dropout: 0.0
+  action_model_lora_rank: 16
+  action_model_lora_target_modules:
+  - vl_qkv
+  - vl_proj
+  - sa_qkv
+  - sa_proj
+  - p_qkv
+  - p_proj
+  - linear1
+  - linear2
+  action_model_use_lora: false
+  add_cross_attention: false
+  architectures: null
+  backbone_lora_alpha: 32
+  backbone_lora_dropout: 0.0
+  backbone_lora_num_layers: -1
+  backbone_lora_rank: 16
+  backbone_lora_target_modules:
+  - q_proj
+  - k_proj
+  - v_proj
+  - o_proj
+  - gate_proj
+  - up_proj
+  - down_proj
+  backbone_use_lora: false
+  bad_words_ids: null
+  begin_suppress_tokens: null
+  bos_token_id: null
+  chunk_size_feed_forward: 0
+  color_jitter_params:
+    brightness: 0.3
+    contrast: 0.4
+    hue: 0.08
+    saturation: 0.5
+  conversation_image_first: false
+  cross_attention_hidden_size: null
+  decoder_start_token_id: null
+  diffusion_model_cfg:
+    action_model_max_seq_len: 512
+    attention_head_dim: 64
+    depth_multi_stream: 4
+    depth_single_stream: 8
+    dropout: 0.2
+    final_dropout: true
+    num_attention_heads: 24
+    output_dim: 1024
+    positional_embeddings: rope_sa_only
+    pre_norm: layer_norm
+    qk_norm: rms_norm
+    rope_theta: 10000.0
+    temb_type: input_token
+    use_swiglu: true
+  diversity_penalty: 0.0
+  do_sample: false
+  dtype: null
+  early_stopping: false
+  encoder_no_repeat_ngram_size: 0
+  eos_token_id: null
+  exponential_decay_length_penalty: null
+  finetuning_task: null
+  forced_bos_token_id: null
+  forced_eos_token_id: null
+  freeze_cog_tokens: false
+  general_embodiment_train_ratio: 0.02
+  id2label:
+    0: LABEL_0
+    1: LABEL_1
+  image_max_area: 65536
+  image_resize_m: 32
+  is_decoder: false
+  is_encoder_decoder: false
+  label2id:
+    LABEL_0: 0
+    LABEL_1: 1
+  length_penalty: 1.0
+  max_length: 20
+  memory_cfg:
+    hidden_size: 4096
+    intermediate_size: 16384
+    max_position_embeddings: 32
+    num_attention_heads: 16
+    num_hidden_layers: 2
+    num_key_value_heads: 16
+    rms_norm_eps: 1.0e-05
+    use_causal_attn: true
+    use_rope: true
+  min_length: 0
+  model_name: /data/taeyoung/models/RLDX-1-VLM
+  model_revision: null
+  model_type: RLDX-1
+  n_cog_tokens: 64
+  no_repeat_ngram_size: 0
+  num_beam_groups: 1
+  num_beams: 1
+  num_return_sequences: 1
+  output_hidden_states: false
+  output_scores: false
+  pad_token_id: null
+  physics_dims: []
+  physics_keys: []
+  prefix: null
+  problem_type: null
+  pruned_heads: {}
+  random_crop_fraction: null
+  random_rotation_angle: null
+  remove_invalid_values: false
+  repetition_penalty: 1.0
+  return_dict: true
+  return_dict_in_generate: false
+  rtc_inference_delay: 0
+  rtc_inference_exec_horizon: 0
+  rtc_inference_mode: none
+  rtc_jacobian_steps_only: 3
+  rtc_training_max_delay: 0
+  sep_token_id: null
+  state_dropout_prob: 0.0
+  suppress_tokens: null
+  task_specific_params: null
+  temperature: 1.0
+  tf_legacy_loss: false
+  tie_encoder_decoder: false
+  tie_word_embeddings: true
+  tokenizer_class: null
+  top_k: 50
+  top_p: 1.0
+  torchscript: false
+  transformers_version: null
+  tune_diffusion_model: true
+  tune_llm: false
+  tune_projector: true
+  tune_top_llm_layers: 4
+  tune_visual: false
+  typical_p: 1.0
+  use_bfloat16: false
+  use_memory: false
+  use_motion: false
+  use_physics: false
+  use_video: true
+  video_length: 1
+  video_stride: 2
+training: !!python/object:rldx.configs.training.training_config.TrainingConfig
+  add_rl_callback: false
+  assert_loss_less_than: null
+  batch_size: null
+  bf16: true
+  dataloader_num_workers: 40
+  ddp_bucket_cap_mb: 100
+  deepspeed_stage: 2
+  enable_open_loop_eval: false
+  enable_profiling: false
+  eval_batch_size: 2
+  eval_bf16: true
+  eval_set_split_ratio: 0.1
+  eval_steps: 500
+  eval_strategy: 'no'
+  experiment_name: rldx1_image_pt_mixv5_bsz8192_100k
+  fp16: false
+  global_batch_size: 4096
+  gradient_accumulation_steps: 2
+  gradient_checkpointing: false
+  learning_rate: 0.0001
+  logging_steps: 10
+  lr_scheduler_type: constant_with_warmup
+  max_concurrent_uploads: 2
+  max_grad_norm: 1.0
+  max_retries: 3
+  max_steps: 100000
+  new_param_warmup_steps: 0
+  num_gpus: 64
+  open_loop_eval_plot_indices: null
+  open_loop_eval_steps_per_traj: 100
+  open_loop_eval_traj_ids:
+  - 0
+  optim: adamw_torch_fused
+  output_dir: /data/taeyoung/workspace/RLDX-1/ckpt/rldx1/pretrained/rldx1_image_pt_mixv5_bsz8192_100k
+  remove_unused_columns: false
+  save_best_eval_metric_greater_is_better: true
+  save_best_eval_metric_name: ''
+  save_steps: 1000
+  save_total_limit: 40
+  save_vl_model: false
+  start_from_checkpoint: null
+  tf32: true
+  transformers_access_token: null
+  transformers_cache_dir: null
+  transformers_local_files_only: false
+  transformers_trust_remote_code: true
+  upload_checkpoints: false
+  upload_every: 1000
+  upload_last_n_checkpoints: 5
+  use_ddp: false
+  use_wandb: true
+  wandb_project: rldx-pretrain
+  warmup_ratio: 0.05
+  warmup_steps: 0
+  weight_decay: 1.0e-05

experiment_cfg/final_model_config.json

@@ -0,0 +1,153 @@
+{
+  "model_type": "RLDX-1",
+  "model_dtype": "bfloat16",
+  "model_name": "/data/taeyoung/models/RLDX-1-VLM",
+  "backbone_model_type": "vtc_qwen3_vl",
+  "model_revision": null,
+  "tune_top_llm_layers": 4,
+  "backbone_embedding_dim": 4096,
+  "tune_llm": false,
+  "tune_visual": false,
+  "select_layer": 18,
+  "reproject_vision": false,
+  "use_flash_attention": true,
+  "load_bf16": true,
+  "backbone_trainable_params_fp32": true,
+  "freeze_cog_tokens": false,
+  "image_max_area": 65536,
+  "image_resize_m": 32,
+  "apply_sincos_state_encoding": false,
+  "use_relative_action": true,
+  "use_percentiles": true,
+  "conversation_image_first": false,
+  "max_state_dim": 64,
+  "max_action_dim": 64,
+  "action_horizon": 16,
+  "hidden_size": 1024,
+  "input_embedding_dim": 1536,
+  "general_embodiment_train_ratio": 0.02,
+  "add_pos_embed": true,
+  "attn_dropout": 0.2,
+  "use_vlln": false,
+  "max_seq_len": 1024,
+  "n_cog_tokens": 64,
+  "diffusion_model_cfg": {
+    "attention_head_dim": 64,
+    "depth_multi_stream": 4,
+    "depth_single_stream": 8,
+    "dropout": 0.2,
+    "final_dropout": true,
+    "num_attention_heads": 24,
+    "output_dim": 1024,
+    "positional_embeddings": "rope_sa_only",
+    "rope_theta": 10000.0,
+    "temb_type": "input_token",
+    "use_swiglu": true,
+    "action_model_max_seq_len": 512,
+    "pre_norm": "layer_norm",
+    "qk_norm": "rms_norm",
+    "sa_dim": 1536,
+    "vl_dim": 4096,
+    "use_physics": false,
+    "physics_dim": 0
+  },
+  "num_inference_timesteps": 4,
+  "noise_beta_alpha": 1.5,
+  "noise_beta_beta": 1.0,
+  "noise_s": 0.999,
+  "num_timestep_buckets": 1000,
+  "tune_projector": true,
+  "tune_diffusion_model": true,
+  "tune_vlln": true,
+  "action_model_use_lora": false,
+  "action_model_lora_rank": 16,
+  "action_model_lora_alpha": 32,
+  "action_model_lora_dropout": 0.0,
+  "action_model_lora_target_modules": [
+    "vl_qkv",
+    "vl_proj",
+    "sa_qkv",
+    "sa_proj",
+    "p_qkv",
+    "p_proj",
+    "linear1",
+    "linear2"
+  ],
+  "backbone_use_lora": false,
+  "backbone_lora_rank": 16,
+  "backbone_lora_alpha": 32,
+  "backbone_lora_dropout": 0.0,
+  "backbone_lora_num_layers": -1,
+  "backbone_lora_target_modules": [
+    "q_proj",
+    "k_proj",
+    "v_proj",
+    "o_proj",
+    "gate_proj",
+    "up_proj",
+    "down_proj"
+  ],
+  "state_dropout_prob": 0.0,
+  "state_additive_noise_scale": 0.0,
+  "rtc_training_max_delay": 0,
+  "rtc_inference_mode": "none",
+  "rtc_inference_delay": 0,
+  "rtc_inference_exec_horizon": 0,
+  "rtc_jacobian_steps_only": 3,
+  "use_memory": false,
+  "memory_length": 4,
+  "memory_n_cog_tokens": null,
+  "concat_memory": false,
+  "memory_dropout_prob": 0.0,
+  "memory_stride": 16,
+  "memory_cfg": {
+    "hidden_size": 4096,
+    "intermediate_size": 16384,
+    "num_hidden_layers": 2,
+    "num_attention_heads": 16,
+    "num_key_value_heads": 16,
+    "max_position_embeddings": 32,
+    "rms_norm_eps": 1e-05,
+    "use_causal_attn": true,
+    "use_rope": true
+  },
+  "use_motion": false,
+  "motion_insert_layer": 9,
+  "motion_d_hid": 512,
+  "motion_window": [
+    5,
+    9,
+    9
+  ],
+  "motion_ext_chnls": [
+    256
+  ],
+  "motion_int_chnls": [
+    256,
+    256,
+    512
+  ],
+  "motion_corr_func": "cosine",
+  "motion_n_encoders": 1,
+  "motion_use_layerscale": false,
+  "motion_layerscale_init": 1e-05,
+  "motion_use_layernorm": false,
+  "motion_use_syncbn": false,
+  "motion_injection_point": "vision_encoder",
+  "motion_pool_type": "avg",
+  "motion_drop": true,
+  "motion_gradient_check": false,
+  "motion_int_mode": "lite",
+  "use_video": true,
+  "video_length": 1,
+  "video_stride": 2,
+  "use_physics": false,
+  "physics_keys": [],
+  "physics_dims": [],
+  "physics_loss_weight": 0.1,
+  "allow_missing_physics": false,
+  "physics_delta_indices": null,
+  "physics_use_flow_matching": true,
+  "physics_dropout_prob": 0.0,
+  "max_num_embodiments": 36
+}

model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43e94d619d95fdac2559c90c232c1c235e626f058903fd181330a1a9efbf84d3
+size 4912540968

model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f24e451e3e3e4323c3a6a312c35d6f788ad714d9e0bab2fef6f69cdea59d855
+size 4446192352

model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a46476ee80d5a18cc2e952e5bc81de530a5f6a7d627850a006a1addf59b95b9
+size 4467155576

model.safetensors.index.json

@@ -0,0 +1,730 @@
+{
+  "metadata": {
+    "total_parameters": 6912894784,
+    "total_size": 13825789568
+  },
+  "weight_map": {
+    "action_model.action_decoder.layer1.W": "model-00003-of-00003.safetensors",
+    "action_model.action_decoder.layer1.b": "model-00003-of-00003.safetensors",
+    "action_model.action_decoder.layer2.W": "model-00003-of-00003.safetensors",
+    "action_model.action_decoder.layer2.b": "model-00003-of-00003.safetensors",
+    "action_model.action_encoder.W1.W": "model-00003-of-00003.safetensors",
+    "action_model.action_encoder.W1.b": "model-00003-of-00003.safetensors",
+    "action_model.action_encoder.W2.W": "model-00003-of-00003.safetensors",
+    "action_model.action_encoder.W2.b": "model-00003-of-00003.safetensors",
+    "action_model.action_encoder.W3.W": "model-00003-of-00003.safetensors",
+    "action_model.action_encoder.W3.b": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.k_norm_sa.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.k_norm_vl.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.q_norm_sa.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.q_norm_vl.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.sa_mlp.w12.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.sa_mlp.w12.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.sa_mlp.w3.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.sa_mlp.w3.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.sa_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.sa_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.sa_qkv.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.sa_qkv.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.vl_mlp.w12.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.vl_mlp.w12.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.vl_mlp.w3.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.vl_mlp.w3.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.vl_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.vl_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.vl_qkv.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.0.vl_qkv.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.k_norm_sa.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.k_norm_vl.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.q_norm_sa.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.q_norm_vl.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.sa_mlp.w12.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.sa_mlp.w12.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.sa_mlp.w3.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.sa_mlp.w3.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.sa_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.sa_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.sa_qkv.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.sa_qkv.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.vl_mlp.w12.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.vl_mlp.w12.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.vl_mlp.w3.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.vl_mlp.w3.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.vl_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.vl_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.vl_qkv.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.1.vl_qkv.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.k_norm_sa.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.k_norm_vl.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.q_norm_sa.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.q_norm_vl.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.sa_mlp.w12.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.sa_mlp.w12.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.sa_mlp.w3.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.sa_mlp.w3.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.sa_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.sa_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.sa_qkv.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.sa_qkv.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.vl_mlp.w12.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.vl_mlp.w12.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.vl_mlp.w3.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.vl_mlp.w3.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.vl_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.vl_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.vl_qkv.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.2.vl_qkv.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.k_norm_sa.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.k_norm_vl.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.q_norm_sa.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.q_norm_vl.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.sa_mlp.w12.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.sa_mlp.w12.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.sa_mlp.w3.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.sa_mlp.w3.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.sa_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.sa_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.sa_qkv.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.sa_qkv.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.vl_mlp.w12.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.vl_mlp.w12.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.vl_mlp.w3.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.vl_mlp.w3.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.vl_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.vl_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.vl_qkv.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.double_blocks.3.vl_qkv.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.proj_out_1.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.proj_out_1.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.proj_out_2.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.proj_out_2.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.0.k_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.0.linear1.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.0.linear1.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.0.linear2.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.0.linear2.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.0.mlp_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.0.mlp_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.0.q_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.1.k_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.1.linear1.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.1.linear1.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.1.linear2.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.1.linear2.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.1.mlp_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.1.mlp_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.1.q_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.2.k_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.2.linear1.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.2.linear1.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.2.linear2.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.2.linear2.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.2.mlp_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.2.mlp_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.2.q_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.3.k_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.3.linear1.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.3.linear1.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.3.linear2.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.3.linear2.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.3.mlp_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.3.mlp_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.3.q_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.4.k_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.4.linear1.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.4.linear1.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.4.linear2.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.4.linear2.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.4.mlp_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.4.mlp_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.4.q_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.5.k_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.5.linear1.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.5.linear1.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.5.linear2.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.5.linear2.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.5.mlp_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.5.mlp_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.5.q_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.6.k_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.6.linear1.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.6.linear1.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.6.linear2.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.6.linear2.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.6.mlp_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.6.mlp_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.6.q_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.7.k_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.7.linear1.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.7.linear1.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.7.linear2.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.7.linear2.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.7.mlp_proj.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.7.mlp_proj.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.single_blocks.7.q_norm.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.timestep_encoder.timestep_embedder.linear_1.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.timestep_encoder.timestep_embedder.linear_1.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.timestep_encoder.timestep_embedder.linear_2.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.timestep_encoder.timestep_embedder.linear_2.weight": "model-00003-of-00003.safetensors",
+    "action_model.model.vl_proj_to_sa.bias": "model-00003-of-00003.safetensors",
+    "action_model.model.vl_proj_to_sa.weight": "model-00003-of-00003.safetensors",
+    "action_model.position_embedding.weight": "model-00003-of-00003.safetensors",
+    "action_model.state_encoder.layer1.W": "model-00003-of-00003.safetensors",
+    "action_model.state_encoder.layer1.b": "model-00003-of-00003.safetensors",
+    "action_model.state_encoder.layer2.W": "model-00003-of-00003.safetensors",
+    "action_model.state_encoder.layer2.b": "model-00003-of-00003.safetensors",
+    "backbone.cog_emb": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.lm_head.weight": "model-00003-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.embed_tokens.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.0.layer.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.0.layer.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.0.layer.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.0.layer.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.0.layer.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.0.layer.self_attn.k_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.0.layer.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.0.layer.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.0.layer.self_attn.q_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.0.layer.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.0.layer.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.1.layer.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.1.layer.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.1.layer.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.1.layer.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.1.layer.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.1.layer.self_attn.k_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.1.layer.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.1.layer.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.1.layer.self_attn.q_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.1.layer.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.1.layer.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.10.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.10.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.10.layer.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.10.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.10.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.10.layer.self_attn.k_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.10.layer.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.10.layer.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.10.layer.self_attn.q_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.10.layer.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.10.layer.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.11.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.11.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.11.layer.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.11.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.11.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.11.layer.self_attn.k_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.11.layer.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.11.layer.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.11.layer.self_attn.q_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.11.layer.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.11.layer.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.12.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.12.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.12.layer.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.12.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.12.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.12.layer.self_attn.k_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.12.layer.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.12.layer.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.12.layer.self_attn.q_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.12.layer.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.12.layer.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.13.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.13.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.13.layer.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.13.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.13.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.13.layer.self_attn.k_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.13.layer.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.13.layer.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.13.layer.self_attn.q_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.13.layer.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.13.layer.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.14.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.14.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.14.layer.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.14.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.14.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.14.layer.self_attn.k_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.14.layer.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.14.layer.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.14.layer.self_attn.q_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.14.layer.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.14.layer.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.15.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.15.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.15.layer.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.15.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.15.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.15.layer.self_attn.k_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.15.layer.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.15.layer.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.15.layer.self_attn.q_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.15.layer.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.15.layer.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.16.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.16.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.16.layer.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.16.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.16.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.16.layer.self_attn.k_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.16.layer.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.16.layer.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.16.layer.self_attn.q_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.16.layer.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.16.layer.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.17.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.17.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.17.layer.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.17.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.17.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.17.layer.self_attn.k_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.17.layer.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.17.layer.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.17.layer.self_attn.q_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.17.layer.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.17.layer.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.2.layer.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.2.layer.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.2.layer.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.2.layer.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.2.layer.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.2.layer.self_attn.k_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.2.layer.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.2.layer.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.2.layer.self_attn.q_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.2.layer.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.2.layer.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.3.layer.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.3.layer.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.3.layer.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.3.layer.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.3.layer.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.3.layer.self_attn.k_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.3.layer.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.3.layer.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.3.layer.self_attn.q_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.3.layer.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.3.layer.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.4.layer.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.4.layer.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.4.layer.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.4.layer.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.4.layer.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.4.layer.self_attn.k_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.4.layer.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.4.layer.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.4.layer.self_attn.q_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.4.layer.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.4.layer.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.5.layer.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.5.layer.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.5.layer.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.5.layer.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.5.layer.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.5.layer.self_attn.k_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.5.layer.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.5.layer.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.5.layer.self_attn.q_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.5.layer.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.5.layer.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.6.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.6.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.6.layer.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.6.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.6.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.6.layer.self_attn.k_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.6.layer.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.6.layer.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.6.layer.self_attn.q_norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.6.layer.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.6.layer.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.7.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.7.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.7.layer.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.7.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.7.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.7.layer.self_attn.k_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.7.layer.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.7.layer.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.7.layer.self_attn.q_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.7.layer.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.7.layer.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.8.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.8.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.8.layer.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.8.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.8.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.8.layer.self_attn.k_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.8.layer.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.8.layer.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.8.layer.self_attn.q_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.8.layer.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.8.layer.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.9.layer.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.9.layer.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.9.layer.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.9.layer.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.9.layer.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.9.layer.self_attn.k_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.9.layer.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.9.layer.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.9.layer.self_attn.q_norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.9.layer.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.layers.9.layer.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.language_model.norm.weight": "model-00002-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.0.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.1.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.10.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.11.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.12.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.13.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.14.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.15.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.16.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.17.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.18.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.19.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.2.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.20.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.21.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.22.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.23.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.24.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.25.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.26.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.3.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.4.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.5.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.6.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.7.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.8.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.attn.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.attn.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.attn.qkv.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.attn.qkv.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.mlp.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.mlp.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.mlp.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.mlp.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.norm1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.norm1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.norm2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.blocks.9.norm2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.0.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.0.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.0.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.0.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.0.norm.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.0.norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.1.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.1.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.1.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.1.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.1.norm.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.1.norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.2.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.2.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.2.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.2.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.2.norm.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.deepstack_merger_list.2.norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.merger.linear_fc1.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.merger.linear_fc1.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.merger.linear_fc2.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.merger.linear_fc2.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.merger.norm.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.merger.norm.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.patch_embed.proj.bias": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.patch_embed.proj.weight": "model-00001-of-00003.safetensors",
+    "backbone.qwen_model.model.visual.pos_embed.weight": "model-00001-of-00003.safetensors"
+  }
+}

processor/embodiment_id.json

@@ -0,0 +1,38 @@
+{
+  "general_embodiment": 0,
+  "fractal20220817_data": 1,
+  "kuka": 2,
+  "bridge_orig": 3,
+  "taco_play": 4,
+  "jaco_play": 5,
+  "berkeley_cable_routing": 6,
+  "roboturk": 7,
+  "viola": 8,
+  "berkeley_autolab_ur5": 9,
+  "toto": 10,
+  "language_table": 11,
+  "stanford_hydra_dataset_converted_externally_to_rlds": 12,
+  "austin_buds_dataset_converted_externally_to_rlds": 13,
+  "nyu_franka_play_dataset_converted_externally_to_rlds": 14,
+  "furniture_bench_dataset_converted_externally_to_rlds": 15,
+  "ucsd_kitchen_dataset_converted_externally_to_rlds": 16,
+  "austin_sailor_dataset_converted_externally_to_rlds": 17,
+  "austin_sirius_dataset_converted_externally_to_rlds": 18,
+  "dlr_edan_shared_control_converted_externally_to_rlds": 19,
+  "iamlab_cmu_pickup_insert_converted_externally_to_rlds": 20,
+  "utaustin_mutex": 21,
+  "berkeley_fanuc_manipulation": 22,
+  "cmu_stretch": 23,
+  "bc_z": 24,
+  "fmb_dataset": 25,
+  "dobbe": 26,
+  "droid": 27,
+  "agibot_dexhand": 28,
+  "agibot_gripper": 29,
+  "galaxea": 30,
+  "humanoid_everyday_g1": 31,
+  "humanoid_everyday_h1": 32,
+  "action_net": 33,
+  "neural_gr1": 34,
+  "new_embodiment": 35
+}

processor/processor_config.json

@@ -0,0 +1,2878 @@
+{
+  "processor_class": "RLDXProcessor",
+  "processor_kwargs": {
+    "modality_configs": {
+      "agibot_gripper": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist_left",
+            "wrist_right"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "state"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "action"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "dlr_edan_shared_control_converted_externally_to_rlds": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "humanoid_everyday_g1": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "egocentric_resized"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "left_arm",
+            "left_hand",
+            "right_arm",
+            "right_hand"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "left_arm",
+            "left_hand",
+            "right_arm",
+            "right_hand"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "austin_sirius_dataset_converted_externally_to_rlds": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "jaco_play": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "cmu_stretch": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "fractal20220817_data": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "berkeley_autolab_ur5": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "language_table": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "viola": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "joint_position",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "furniture_bench_dataset_converted_externally_to_rlds": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "agibot_dexhand": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "state"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "action"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "taco_play": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "berkeley_fanuc_manipulation": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "joint_position",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "bc_z": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "toto": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "joint_position",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "dobbe": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "droid": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "secondary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "ucsd_kitchen_dataset_converted_externally_to_rlds": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "joint_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "stanford_hydra_dataset_converted_externally_to_rlds": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "iamlab_cmu_pickup_insert_converted_externally_to_rlds": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "nyu_franka_play_dataset_converted_externally_to_rlds": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "secondary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "austin_sailor_dataset_converted_externally_to_rlds": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "utaustin_mutex": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "joint_position",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "fmb_dataset": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "secondary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "bridge_orig": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "secondary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "austin_buds_dataset_converted_externally_to_rlds": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "joint_position",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "neural_gr1": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "ego_view"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "left_arm",
+            "left_hand",
+            "left_leg",
+            "neck",
+            "right_arm",
+            "right_hand",
+            "right_leg",
+            "waist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "left_arm",
+            "left_hand",
+            "left_leg",
+            "neck",
+            "right_arm",
+            "right_hand",
+            "right_leg",
+            "waist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "galaxea": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "wrist_left",
+            "wrist_right"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "state"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "action"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "roboturk": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "none"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "berkeley_cable_routing": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary",
+            "secondary",
+            "wrist"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "joint_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "action_net": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "state"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "action"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "kuka": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "primary"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_position"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "end_effector_position",
+            "end_effector_rotation",
+            "gripper_close"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "DELTA",
+              "type": "EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      },
+      "humanoid_everyday_h1": {
+        "video": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "egocentric_resized"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "state": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "left_arm",
+            "left_hand",
+            "right_arm",
+            "right_hand"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        },
+        "action": {
+          "delta_indices": [
+            0,
+            1,
+            2,
+            3,
+            4,
+            5,
+            6,
+            7,
+            8,
+            9,
+            10,
+            11,
+            12,
+            13,
+            14,
+            15
+          ],
+          "modality_keys": [
+            "left_arm",
+            "left_hand",
+            "right_arm",
+            "right_hand"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": [
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            },
+            {
+              "rep": "ABSOLUTE",
+              "type": "NON_EEF",
+              "format": "DEFAULT",
+              "state_key": null
+            }
+          ]
+        },
+        "language": {
+          "delta_indices": [
+            0
+          ],
+          "modality_keys": [
+            "annotation.human.action.task_description"
+          ],
+          "sin_cos_embedding_keys": null,
+          "mean_std_embedding_keys": null,
+          "action_configs": null
+        }
+      }
+    },
+    "image_max_area": 65536,
+    "image_resize_m": 32,
+    "random_crop_fraction": null,
+    "random_rotation_angle": null,
+    "color_jitter_params": {
+      "brightness": 0.3,
+      "contrast": 0.4,
+      "saturation": 0.5,
+      "hue": 0.08
+    },
+    "model_name": "/data/taeyoung/models/RLDX-1-VLM",
+    "model_type": "vtc_qwen3_vl",
+    "formalize_language": true,
+    "max_state_dim": 64,
+    "max_action_dim": 64,
+    "max_action_horizon": 16,
+    "use_percentiles": true,
+    "clip_outliers": true,
+    "apply_sincos_state_encoding": false,
+    "use_relative_action": true,
+    "memory_length": 1,
+    "general_embodiment_train_ratio": 0.02,
+    "allow_missing_physics": false,
+    "physics_keys": null,
+    "physics_dims": null,
+    "conversation_image_first": false
+  }
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37e8ff7d0b1e021327b09e1c7ddaa4170cfac8aa4db96e9aacd1a519ecd567a9
+size 8017

wandb_config.json

@@ -0,0 +1 @@
+{"project": "rldx-pretrain", "run_id": "rldx1_image_pt_mixv5_bsz8192_100k"}

zero_to_fp32.py

@@ -0,0 +1,760 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example:
+#   python zero_to_fp32.py . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import gc
+import json
+import numpy as np
+from tqdm import tqdm
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device, weights_only=False)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+    total_files = len(files)
+    state_dicts = []
+    for f in tqdm(files, desc='Loading checkpoint shards'):
+        state_dict = torch.load(f, map_location=device, mmap=True, weights_only=False)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if ZERO_STAGE not in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+class GatheredTensor:
+    """
+    A pseudo tensor that collects partitioned weights.
+    It is more memory efficient when there are multiple groups.
+    """
+
+    def __init__(self, flat_groups, flat_groups_offset, offset, partitioned_numel, shape):
+        self.flat_groups = flat_groups
+        self.flat_groups_offset = flat_groups_offset
+        self.offset = offset
+        self.partitioned_numel = partitioned_numel
+        self.shape = shape
+        self.dtype = self.flat_groups[0][0].dtype
+
+    def contiguous(self):
+        """
+        Merge partitioned weights from flat_groups into a single tensor.
+        """
+        end_idx = self.offset + self.partitioned_numel
+        world_size = len(self.flat_groups)
+        pad_flat_param_chunks = []
+
+        for rank_i in range(world_size):
+            # for each rank, we need to collect weights from related group/groups
+            flat_groups_at_rank_i = self.flat_groups[rank_i]
+            start_group_id = None
+            end_group_id = None
+            for group_id in range(len(self.flat_groups_offset)):
+                if self.flat_groups_offset[group_id] <= self.offset < self.flat_groups_offset[group_id + 1]:
+                    start_group_id = group_id
+                if self.flat_groups_offset[group_id] < end_idx <= self.flat_groups_offset[group_id + 1]:
+                    end_group_id = group_id
+                    break
+            # collect weights from related group/groups
+            for group_id in range(start_group_id, end_group_id + 1):
+                flat_tensor = flat_groups_at_rank_i[group_id]
+                start_offset = self.offset - self.flat_groups_offset[group_id]
+                end_offset = min(end_idx, self.flat_groups_offset[group_id + 1]) - self.flat_groups_offset[group_id]
+                pad_flat_param_chunks.append(flat_tensor[start_offset:end_offset])
+
+        # collect weights from all ranks
+        pad_flat_param = torch.cat(pad_flat_param_chunks, dim=0)
+        param = pad_flat_param[:self.shape.numel()].view(self.shape).contiguous()
+        return param
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = sum([flat_group.numel() for flat_group in fp32_flat_groups[0]]) * world_size
+
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    flat_groups_offset = [0] + list(np.cumsum([flat_tensor.numel() for flat_tensor in fp32_flat_groups[0]]))
+    for name, shape in tqdm(param_shapes.items(), desc='Gathering sharded weights'):
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # memory efficient tensor
+        tensor = GatheredTensor(fp32_flat_groups, flat_groups_offset, offset, partitioned_numel, shape)
+        state_dict[name] = tensor
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def to_torch_tensor(state_dict, return_empty_tensor=False):
+    """
+    Convert state_dict of GatheredTensor to torch tensor
+    """
+    torch_state_dict = {}
+    converted_tensors = {}
+    for name, tensor in state_dict.items():
+        tensor_id = id(tensor)
+        if tensor_id in converted_tensors:  # shared tensors
+            shared_tensor = torch_state_dict[converted_tensors[tensor_id]]
+            torch_state_dict[name] = shared_tensor
+        else:
+            converted_tensors[tensor_id] = name
+            if return_empty_tensor:
+                torch_state_dict[name] = torch.empty(tensor.shape, dtype=tensor.dtype)
+            else:
+                torch_state_dict[name] = tensor.contiguous()
+    return torch_state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                             tag=None,
+                                             exclude_frozen_parameters=False,
+                                             lazy_mode=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+        - ``lazy_mode``: get state_dict in lazy mode. It returns a dict of pesduo tensor instead of torch tensor, which is more memory efficient.
+          Convert the pesduo tensor to torch tensor by ``.contiguous()``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    Note: the above usage may not work if your application doesn't have sufficient free CPU memory.
+    You may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint. Or you can load state_dict in lazy mode ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, lazy_mode=True) # not on cpu
+        for name, lazy_tensor in state_dict.item():
+            tensor = lazy_tensor.contiguous()  # to cpu
+            print(name, tensor)
+            # del tensor to release memory if it no longer in use
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    state_dict = _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+    if lazy_mode:
+        return state_dict
+    else:
+        return to_torch_tensor(state_dict)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir,
+                                               output_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               tag=None,
+                                               exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir,
+                                                          tag,
+                                                          exclude_frozen_parameters,
+                                                          lazy_mode=True)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        # an memory-efficient approach for sharding
+        empty_state_dict = to_torch_tensor(state_dict, return_empty_tensor=True)
+        state_dict_split = split_torch_state_dict_into_shards(empty_state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model by shard
+    os.makedirs(output_dir, exist_ok=True)
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard_state_dict = {tensor_name: state_dict[tensor_name] for tensor_name in tensors}
+        shard_state_dict = to_torch_tensor(shard_state_dict)
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard_state_dict, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard_state_dict, output_path)
+        # release the memory of current shard
+        for tensor_name in list(shard_state_dict.keys()):
+            del state_dict[tensor_name]
+            del shard_state_dict[tensor_name]
+        del shard_state_dict
+        gc.collect()
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info("Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info("Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument("output_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)