Upload folder using huggingface_hub
Browse files- *.pth/config.json +90 -0
- *.pth/embodiment_id.json +55 -0
- *.pth/experiment_cfg/conf.yaml +221 -0
- *.pth/experiment_cfg/config.yaml +248 -0
- *.pth/experiment_cfg/dataset_statistics.json +321 -0
- *.pth/experiment_cfg/final_model_config.json +56 -0
- *.pth/experiment_cfg/final_processor_config.json +0 -0
- *.pth/model-00001-of-00003.safetensors +3 -0
- *.pth/model-00002-of-00003.safetensors +3 -0
- *.pth/model-00003-of-00003.safetensors +3 -0
- *.pth/model.safetensors.index.json +0 -0
- *.pth/processor_config.json +1132 -0
- *.pth/statistics.json +0 -0
- *.pth/trainer_state.json +0 -0
- *.pth/training_args.bin +3 -0
- *.pth/wandb_config.json +1 -0
*.pth/config.json
ADDED
|
@@ -0,0 +1,90 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"action_horizon": 40,
|
| 3 |
+
"add_pos_embed": true,
|
| 4 |
+
"apply_sincos_state_encoding": false,
|
| 5 |
+
"architectures": [
|
| 6 |
+
"Gr00tN1d7"
|
| 7 |
+
],
|
| 8 |
+
"attn_dropout": 0.2,
|
| 9 |
+
"attn_implementation": null,
|
| 10 |
+
"backbone_embedding_dim": 2048,
|
| 11 |
+
"backbone_trainable_params_fp32": true,
|
| 12 |
+
"clip_outliers": true,
|
| 13 |
+
"color_jitter_params": {
|
| 14 |
+
"brightness": 0.3,
|
| 15 |
+
"contrast": 0.4,
|
| 16 |
+
"hue": 0.08,
|
| 17 |
+
"saturation": 0.5
|
| 18 |
+
},
|
| 19 |
+
"crop_fraction": 0.95,
|
| 20 |
+
"diffusion_model_cfg": {
|
| 21 |
+
"attention_head_dim": 48,
|
| 22 |
+
"dropout": 0.2,
|
| 23 |
+
"final_dropout": true,
|
| 24 |
+
"interleave_self_attention": true,
|
| 25 |
+
"norm_type": "ada_norm",
|
| 26 |
+
"num_attention_heads": 32,
|
| 27 |
+
"num_layers": 32,
|
| 28 |
+
"output_dim": 1024,
|
| 29 |
+
"positional_embeddings": null
|
| 30 |
+
},
|
| 31 |
+
"dtype": "float32",
|
| 32 |
+
"exclude_state": false,
|
| 33 |
+
"formalize_language": true,
|
| 34 |
+
"hidden_size": 1024,
|
| 35 |
+
"image_crop_size": [
|
| 36 |
+
230,
|
| 37 |
+
230
|
| 38 |
+
],
|
| 39 |
+
"image_target_size": [
|
| 40 |
+
256,
|
| 41 |
+
256
|
| 42 |
+
],
|
| 43 |
+
"letter_box_transform": false,
|
| 44 |
+
"load_bf16": false,
|
| 45 |
+
"max_action_dim": 132,
|
| 46 |
+
"max_num_embodiments": 32,
|
| 47 |
+
"max_seq_len": 1024,
|
| 48 |
+
"max_state_dim": 132,
|
| 49 |
+
"model_dtype": "bfloat16",
|
| 50 |
+
"model_name": "nvidia/Cosmos-Reason2-2B",
|
| 51 |
+
"model_type": "Gr00tN1d7",
|
| 52 |
+
"noise_beta_alpha": 1.5,
|
| 53 |
+
"noise_beta_beta": 1.0,
|
| 54 |
+
"noise_s": 0.999,
|
| 55 |
+
"num_inference_timesteps": 4,
|
| 56 |
+
"num_timestep_buckets": 1000,
|
| 57 |
+
"random_history_crop": true,
|
| 58 |
+
"random_rotation_angle": 0,
|
| 59 |
+
"reproject_vision": false,
|
| 60 |
+
"rtc_ramp_rate": 6.0,
|
| 61 |
+
"select_layer": 16,
|
| 62 |
+
"shortest_image_edge": 256,
|
| 63 |
+
"state_dropout_prob": 0.2,
|
| 64 |
+
"state_gaussian_noise_std": 0.0,
|
| 65 |
+
"transformers_version": "4.57.3",
|
| 66 |
+
"tune_diffusion_model": true,
|
| 67 |
+
"tune_linear": true,
|
| 68 |
+
"tune_llm": false,
|
| 69 |
+
"tune_projector": true,
|
| 70 |
+
"tune_top_llm_layers": 0,
|
| 71 |
+
"tune_visual": false,
|
| 72 |
+
"tune_vlln": true,
|
| 73 |
+
"use_albumentations": true,
|
| 74 |
+
"use_alternate_vl_dit": true,
|
| 75 |
+
"use_flash_attention": true,
|
| 76 |
+
"use_future_tokens": false,
|
| 77 |
+
"use_mean_std": false,
|
| 78 |
+
"use_percentiles": false,
|
| 79 |
+
"use_relative_action": true,
|
| 80 |
+
"use_vl_self_attention": true,
|
| 81 |
+
"use_vlln": true,
|
| 82 |
+
"vl_self_attention_cfg": {
|
| 83 |
+
"attention_head_dim": 64,
|
| 84 |
+
"dropout": 0.2,
|
| 85 |
+
"final_dropout": true,
|
| 86 |
+
"num_attention_heads": 32,
|
| 87 |
+
"num_layers": 4,
|
| 88 |
+
"positional_embeddings": null
|
| 89 |
+
}
|
| 90 |
+
}
|
*.pth/embodiment_id.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"robocasa_panda_omron": 13,
|
| 3 |
+
"oxe_droid": 17,
|
| 4 |
+
"oxe_fractal": 18,
|
| 5 |
+
"oxe_language_table": 19,
|
| 6 |
+
"oxe_bridge": 20,
|
| 7 |
+
"unknown": 22,
|
| 8 |
+
"gr1_unified": 20,
|
| 9 |
+
"agibot": 26,
|
| 10 |
+
"sim_behavior_r1_pro": 23,
|
| 11 |
+
"xdof": 24,
|
| 12 |
+
"xdof_oss_data": 25,
|
| 13 |
+
"unitree_g1_full_body_with_waist_height_nav_cmd": 25,
|
| 14 |
+
"real_r1_pro_sharpa": 27,
|
| 15 |
+
"real_r1_pro_sharpa_add_view": 27,
|
| 16 |
+
"real_r1_pro_sharpa_relative_arm_joint": 26,
|
| 17 |
+
"real_r1_pro_sharpa_delta_eef": 26,
|
| 18 |
+
"real_r1_pro_sharpa_absolute_eef": 26,
|
| 19 |
+
"real_r1_pro_sharpa_meanstd": 26,
|
| 20 |
+
"real_r1_pro_sharpa_relative_eef": 26,
|
| 21 |
+
"real_r1_pro_sharpa_relative_eef_add_view": 26,
|
| 22 |
+
"real_r1_pro_sharpa_relative_eef_relative_hand": 26,
|
| 23 |
+
"real_r1_pro_sharpa_relative_eef_human": 26,
|
| 24 |
+
"real_r1_pro_sharpa_relative_eef_human_add_view": 26,
|
| 25 |
+
"real_r1_pro_sharpa_relative_eef_human_relative_hand": 26,
|
| 26 |
+
"real_r1_pro_sharpa_relative_eef_egodex": 26,
|
| 27 |
+
"real_r1_pro_sharpa_relative_eef_egodex_relative_hand": 26,
|
| 28 |
+
"real_r1_pro_sharpa_relative_eef_egodex_wrist_only": 26,
|
| 29 |
+
"real_r1_pro_sharpa_relative_eef_maxinsights": 26,
|
| 30 |
+
"real_r1_pro_sharpa_relative_eef_maxinsights_relative_hand": 26,
|
| 31 |
+
"real_r1_pro_sharpa_relative_eef_mecka": 26,
|
| 32 |
+
"real_r1_pro_sharpa_relative_eef_mecka_relative_hand": 26,
|
| 33 |
+
"real_g1_relative_eef_absolute_joints": 25,
|
| 34 |
+
"real_g1_relative_eef_absolute_joints_wrist_cam": 25,
|
| 35 |
+
"real_g1_relative_eef_relative_joints": 25,
|
| 36 |
+
"real_r1_pro_sharpa_relative_eef_relative_hand_relative_joint": 26,
|
| 37 |
+
"real_r1_pro_sharpa_relative_joint": 29,
|
| 38 |
+
"oxe_droid_relative_eef_relative_joint": 24,
|
| 39 |
+
"oxe_droid_relative_eef_relative_joint_swapped": 24,
|
| 40 |
+
"oxe_droid_relative_eef_relative_joint_upweight_z": 24,
|
| 41 |
+
"oxe_droid_relative_eef_relative_joint_upweight_z_swapped": 24,
|
| 42 |
+
"oxe_droid_relative_eef_relative_joint_3view": 24,
|
| 43 |
+
"oxe_droid_relative_eef_relative_joint_3view_swapped": 24,
|
| 44 |
+
"oxe_droid_relative_eef": 24,
|
| 45 |
+
"oxe_droid_joint_position_relative": 24,
|
| 46 |
+
"xdof_relative_eef_relative_joint": 27,
|
| 47 |
+
"xdof_relative_eef_relative_joint_subtask": 27,
|
| 48 |
+
"xdof_relative_eef": 27,
|
| 49 |
+
"xdof_relative_joint": 28,
|
| 50 |
+
"simpler_env_google": 0,
|
| 51 |
+
"simpler_env_widowx": 1,
|
| 52 |
+
"libero_sim": 2,
|
| 53 |
+
"droid_sim": 3,
|
| 54 |
+
"new_embodiment": 10
|
| 55 |
+
}
|
*.pth/experiment_cfg/conf.yaml
ADDED
|
@@ -0,0 +1,221 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
load_config_path: null
|
| 2 |
+
model:
|
| 3 |
+
model_type: Gr00tN1d7
|
| 4 |
+
model_dtype: bfloat16
|
| 5 |
+
model_name: nvidia/Cosmos-Reason2-2B
|
| 6 |
+
backbone_model_type: qwen
|
| 7 |
+
model_revision: null
|
| 8 |
+
tune_top_llm_layers: 0
|
| 9 |
+
backbone_embedding_dim: 2048
|
| 10 |
+
tune_llm: false
|
| 11 |
+
tune_visual: false
|
| 12 |
+
select_layer: 12
|
| 13 |
+
reproject_vision: false
|
| 14 |
+
use_flash_attention: true
|
| 15 |
+
load_bf16: false
|
| 16 |
+
backbone_trainable_params_fp32: true
|
| 17 |
+
image_crop_size:
|
| 18 |
+
- 230
|
| 19 |
+
- 230
|
| 20 |
+
image_target_size:
|
| 21 |
+
- 256
|
| 22 |
+
- 256
|
| 23 |
+
shortest_image_edge: null
|
| 24 |
+
crop_fraction: null
|
| 25 |
+
random_rotation_angle: null
|
| 26 |
+
color_jitter_params: null
|
| 27 |
+
use_albumentations_transforms: true
|
| 28 |
+
extra_augmentation_config: null
|
| 29 |
+
formalize_language: true
|
| 30 |
+
apply_sincos_state_encoding: false
|
| 31 |
+
use_percentiles: false
|
| 32 |
+
clip_outliers: true
|
| 33 |
+
use_relative_action: true
|
| 34 |
+
max_state_dim: 132
|
| 35 |
+
max_action_dim: 132
|
| 36 |
+
action_horizon: 40
|
| 37 |
+
hidden_size: 1024
|
| 38 |
+
input_embedding_dim: 1536
|
| 39 |
+
state_history_length: 1
|
| 40 |
+
add_pos_embed: true
|
| 41 |
+
attn_dropout: 0.2
|
| 42 |
+
use_vlln: true
|
| 43 |
+
max_seq_len: 1024
|
| 44 |
+
use_alternate_vl_dit: true
|
| 45 |
+
attend_text_every_n_blocks: 2
|
| 46 |
+
diffusion_model_cfg:
|
| 47 |
+
positional_embeddings: null
|
| 48 |
+
num_layers: 16
|
| 49 |
+
num_attention_heads: 32
|
| 50 |
+
attention_head_dim: 48
|
| 51 |
+
norm_type: ada_norm
|
| 52 |
+
dropout: 0.2
|
| 53 |
+
final_dropout: true
|
| 54 |
+
output_dim: 1024
|
| 55 |
+
interleave_self_attention: true
|
| 56 |
+
num_inference_timesteps: 4
|
| 57 |
+
noise_beta_alpha: 1.5
|
| 58 |
+
noise_beta_beta: 1.0
|
| 59 |
+
noise_s: 0.999
|
| 60 |
+
num_timestep_buckets: 1000
|
| 61 |
+
tune_projector: true
|
| 62 |
+
tune_diffusion_model: true
|
| 63 |
+
tune_vlln: true
|
| 64 |
+
state_dropout_prob: 0.2
|
| 65 |
+
exclude_state: false
|
| 66 |
+
use_mean_std: false
|
| 67 |
+
max_num_embodiments: 32
|
| 68 |
+
data:
|
| 69 |
+
datasets:
|
| 70 |
+
- dataset_paths:
|
| 71 |
+
- /data/dataset
|
| 72 |
+
embodiment_tag: new_embodiment
|
| 73 |
+
mix_ratio: 1.0
|
| 74 |
+
dataset_type: physical_embodiment
|
| 75 |
+
val_dataset_path: null
|
| 76 |
+
modality_configs:
|
| 77 |
+
new_embodiment:
|
| 78 |
+
video:
|
| 79 |
+
delta_indices:
|
| 80 |
+
- 0
|
| 81 |
+
modality_keys:
|
| 82 |
+
- cam_left_head
|
| 83 |
+
- cam_left_wrist
|
| 84 |
+
- cam_right_wrist
|
| 85 |
+
sin_cos_embedding_keys: null
|
| 86 |
+
mean_std_embedding_keys: null
|
| 87 |
+
action_configs: null
|
| 88 |
+
state:
|
| 89 |
+
delta_indices:
|
| 90 |
+
- 0
|
| 91 |
+
modality_keys:
|
| 92 |
+
- arm_left
|
| 93 |
+
- arm_right
|
| 94 |
+
- odometry
|
| 95 |
+
sin_cos_embedding_keys: null
|
| 96 |
+
mean_std_embedding_keys: null
|
| 97 |
+
action_configs: null
|
| 98 |
+
action:
|
| 99 |
+
delta_indices:
|
| 100 |
+
- 0
|
| 101 |
+
- 1
|
| 102 |
+
- 2
|
| 103 |
+
- 3
|
| 104 |
+
- 4
|
| 105 |
+
- 5
|
| 106 |
+
- 6
|
| 107 |
+
- 7
|
| 108 |
+
- 8
|
| 109 |
+
- 9
|
| 110 |
+
- 10
|
| 111 |
+
- 11
|
| 112 |
+
- 12
|
| 113 |
+
- 13
|
| 114 |
+
- 14
|
| 115 |
+
- 15
|
| 116 |
+
modality_keys:
|
| 117 |
+
- arm_left
|
| 118 |
+
- arm_right
|
| 119 |
+
- odometry
|
| 120 |
+
sin_cos_embedding_keys: null
|
| 121 |
+
mean_std_embedding_keys: null
|
| 122 |
+
action_configs:
|
| 123 |
+
- rep: ABSOLUTE
|
| 124 |
+
type: NON_EEF
|
| 125 |
+
format: DEFAULT
|
| 126 |
+
state_key: null
|
| 127 |
+
- rep: ABSOLUTE
|
| 128 |
+
type: NON_EEF
|
| 129 |
+
format: DEFAULT
|
| 130 |
+
state_key: null
|
| 131 |
+
- rep: ABSOLUTE
|
| 132 |
+
type: NON_EEF
|
| 133 |
+
format: DEFAULT
|
| 134 |
+
state_key: null
|
| 135 |
+
language:
|
| 136 |
+
delta_indices:
|
| 137 |
+
- 0
|
| 138 |
+
modality_keys:
|
| 139 |
+
- annotation.human.task_description
|
| 140 |
+
sin_cos_embedding_keys: null
|
| 141 |
+
mean_std_embedding_keys: null
|
| 142 |
+
action_configs: null
|
| 143 |
+
download_cache: false
|
| 144 |
+
shard_size: 1024
|
| 145 |
+
episode_sampling_rate: 0.1
|
| 146 |
+
num_shards_per_epoch: 100000
|
| 147 |
+
override_pretraining_statistics: true
|
| 148 |
+
mode: single_turn
|
| 149 |
+
random_chop: 0.0
|
| 150 |
+
mock_dataset_mode: false
|
| 151 |
+
shuffle: true
|
| 152 |
+
seed: 42
|
| 153 |
+
multiprocessing_context: fork
|
| 154 |
+
allow_padding: false
|
| 155 |
+
subsample_ratio: 1.0
|
| 156 |
+
image_crop_size:
|
| 157 |
+
- 244
|
| 158 |
+
- 244
|
| 159 |
+
image_target_size:
|
| 160 |
+
- 224
|
| 161 |
+
- 224
|
| 162 |
+
video_backend: torchcodec
|
| 163 |
+
training:
|
| 164 |
+
output_dir: /data/checkpoints/task0025_subtask_lang_final_layer_vlm_threecam_minmax_80k_gpu1_20260618_045240
|
| 165 |
+
experiment_name: task0025_subtask_lang_final_layer_vlm_threecam_minmax_80k_gpu1_20260618_045240
|
| 166 |
+
max_steps: 80000
|
| 167 |
+
global_batch_size: 8
|
| 168 |
+
batch_size: null
|
| 169 |
+
gradient_accumulation_steps: 1
|
| 170 |
+
learning_rate: 0.0001
|
| 171 |
+
lr_scheduler_type: cosine
|
| 172 |
+
weight_decay: 1.0e-05
|
| 173 |
+
warmup_ratio: 0.05
|
| 174 |
+
warmup_steps: 0
|
| 175 |
+
max_grad_norm: 1.0
|
| 176 |
+
optim: adamw_torch
|
| 177 |
+
start_from_checkpoint: /data/base_model
|
| 178 |
+
skip_weight_loading: false
|
| 179 |
+
tf32: true
|
| 180 |
+
fp16: false
|
| 181 |
+
bf16: true
|
| 182 |
+
eval_bf16: true
|
| 183 |
+
logging_steps: 10
|
| 184 |
+
save_steps: 20000
|
| 185 |
+
save_total_limit: 2
|
| 186 |
+
save_vl_model: false
|
| 187 |
+
save_only_model: true
|
| 188 |
+
upload_checkpoints: false
|
| 189 |
+
upload_every: 1000
|
| 190 |
+
upload_last_n_checkpoints: 5
|
| 191 |
+
max_concurrent_uploads: 2
|
| 192 |
+
eval_strategy: 'no'
|
| 193 |
+
eval_steps: 500
|
| 194 |
+
eval_set_split_ratio: 0.1
|
| 195 |
+
eval_batch_size: 2
|
| 196 |
+
save_best_eval_metric_name: ''
|
| 197 |
+
save_best_eval_metric_greater_is_better: true
|
| 198 |
+
deepspeed_stage: 2
|
| 199 |
+
gradient_checkpointing: false
|
| 200 |
+
transformers_trust_remote_code: true
|
| 201 |
+
transformers_local_files_only: false
|
| 202 |
+
transformers_cache_dir: null
|
| 203 |
+
transformers_access_token: null
|
| 204 |
+
use_ddp: false
|
| 205 |
+
ddp_bucket_cap_mb: 100
|
| 206 |
+
num_gpus: 1
|
| 207 |
+
dataloader_num_workers: 8
|
| 208 |
+
remove_unused_columns: false
|
| 209 |
+
use_wandb: false
|
| 210 |
+
wandb_project: finetune-gr00t-n1d7
|
| 211 |
+
enable_profiling: false
|
| 212 |
+
max_retries: 3
|
| 213 |
+
assert_loss_less_than: null
|
| 214 |
+
add_rl_callback: false
|
| 215 |
+
enable_open_loop_eval: false
|
| 216 |
+
open_loop_eval_traj_ids:
|
| 217 |
+
- 0
|
| 218 |
+
open_loop_eval_steps_per_traj: 100
|
| 219 |
+
open_loop_eval_plot_indices: null
|
| 220 |
+
max_steps: 80000
|
| 221 |
+
save_steps: 20000
|
*.pth/experiment_cfg/config.yaml
ADDED
|
@@ -0,0 +1,248 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
!!python/object:gr00t.configs.base_config.Config
|
| 2 |
+
data: !!python/object:gr00t.configs.data.data_config.DataConfig
|
| 3 |
+
allow_padding: false
|
| 4 |
+
datasets:
|
| 5 |
+
- !!python/object:gr00t.configs.data.data_config.SingleDatasetConfig
|
| 6 |
+
dataset_paths:
|
| 7 |
+
- /data/dataset
|
| 8 |
+
dataset_type: physical_embodiment
|
| 9 |
+
embodiment_tag: new_embodiment
|
| 10 |
+
mix_ratio: 1.0
|
| 11 |
+
val_dataset_path: null
|
| 12 |
+
download_cache: false
|
| 13 |
+
episode_sampling_rate: 0.1
|
| 14 |
+
image_crop_size:
|
| 15 |
+
- 244
|
| 16 |
+
- 244
|
| 17 |
+
image_target_size:
|
| 18 |
+
- 224
|
| 19 |
+
- 224
|
| 20 |
+
mock_dataset_mode: false
|
| 21 |
+
modality_configs:
|
| 22 |
+
new_embodiment:
|
| 23 |
+
action: !!python/object:gr00t.data.types.ModalityConfig
|
| 24 |
+
action_configs:
|
| 25 |
+
- !!python/object:gr00t.data.types.ActionConfig
|
| 26 |
+
format: &id001 !!python/object/apply:gr00t.data.types.ActionFormat
|
| 27 |
+
- default
|
| 28 |
+
rep: &id002 !!python/object/apply:gr00t.data.types.ActionRepresentation
|
| 29 |
+
- absolute
|
| 30 |
+
state_key: null
|
| 31 |
+
type: &id003 !!python/object/apply:gr00t.data.types.ActionType
|
| 32 |
+
- non_eef
|
| 33 |
+
- !!python/object:gr00t.data.types.ActionConfig
|
| 34 |
+
format: *id001
|
| 35 |
+
rep: *id002
|
| 36 |
+
state_key: null
|
| 37 |
+
type: *id003
|
| 38 |
+
- !!python/object:gr00t.data.types.ActionConfig
|
| 39 |
+
format: *id001
|
| 40 |
+
rep: *id002
|
| 41 |
+
state_key: null
|
| 42 |
+
type: *id003
|
| 43 |
+
delta_indices:
|
| 44 |
+
- 0
|
| 45 |
+
- 1
|
| 46 |
+
- 2
|
| 47 |
+
- 3
|
| 48 |
+
- 4
|
| 49 |
+
- 5
|
| 50 |
+
- 6
|
| 51 |
+
- 7
|
| 52 |
+
- 8
|
| 53 |
+
- 9
|
| 54 |
+
- 10
|
| 55 |
+
- 11
|
| 56 |
+
- 12
|
| 57 |
+
- 13
|
| 58 |
+
- 14
|
| 59 |
+
- 15
|
| 60 |
+
mean_std_embedding_keys: null
|
| 61 |
+
modality_keys:
|
| 62 |
+
- arm_left
|
| 63 |
+
- arm_right
|
| 64 |
+
- odometry
|
| 65 |
+
sin_cos_embedding_keys: null
|
| 66 |
+
language: !!python/object:gr00t.data.types.ModalityConfig
|
| 67 |
+
action_configs: null
|
| 68 |
+
delta_indices:
|
| 69 |
+
- 0
|
| 70 |
+
mean_std_embedding_keys: null
|
| 71 |
+
modality_keys:
|
| 72 |
+
- annotation.human.task_description
|
| 73 |
+
sin_cos_embedding_keys: null
|
| 74 |
+
state: !!python/object:gr00t.data.types.ModalityConfig
|
| 75 |
+
action_configs: null
|
| 76 |
+
delta_indices:
|
| 77 |
+
- 0
|
| 78 |
+
mean_std_embedding_keys: null
|
| 79 |
+
modality_keys:
|
| 80 |
+
- arm_left
|
| 81 |
+
- arm_right
|
| 82 |
+
- odometry
|
| 83 |
+
sin_cos_embedding_keys: null
|
| 84 |
+
video: !!python/object:gr00t.data.types.ModalityConfig
|
| 85 |
+
action_configs: null
|
| 86 |
+
delta_indices:
|
| 87 |
+
- 0
|
| 88 |
+
mean_std_embedding_keys: null
|
| 89 |
+
modality_keys:
|
| 90 |
+
- cam_left_head
|
| 91 |
+
- cam_left_wrist
|
| 92 |
+
- cam_right_wrist
|
| 93 |
+
sin_cos_embedding_keys: null
|
| 94 |
+
mode: single_turn
|
| 95 |
+
multiprocessing_context: fork
|
| 96 |
+
num_shards_per_epoch: 100000
|
| 97 |
+
override_pretraining_statistics: true
|
| 98 |
+
random_chop: 0.0
|
| 99 |
+
seed: 42
|
| 100 |
+
shard_size: 1024
|
| 101 |
+
shuffle: true
|
| 102 |
+
subsample_ratio: 1.0
|
| 103 |
+
video_backend: torchcodec
|
| 104 |
+
load_config_path: null
|
| 105 |
+
model: !!python/object:gr00t.configs.model.gr00t_n1d7.Gr00tN1d7Config
|
| 106 |
+
_attn_implementation_internal: null
|
| 107 |
+
_commit_hash: null
|
| 108 |
+
_name_or_path: ''
|
| 109 |
+
_output_attentions: false
|
| 110 |
+
add_cross_attention: false
|
| 111 |
+
architectures: null
|
| 112 |
+
backbone_trainable_params_fp32: true
|
| 113 |
+
bad_words_ids: null
|
| 114 |
+
begin_suppress_tokens: null
|
| 115 |
+
bos_token_id: null
|
| 116 |
+
chunk_size_feed_forward: 0
|
| 117 |
+
clip_outliers: true
|
| 118 |
+
color_jitter_params: null
|
| 119 |
+
cross_attention_hidden_size: null
|
| 120 |
+
decoder_start_token_id: null
|
| 121 |
+
diffusion_model_cfg:
|
| 122 |
+
attention_head_dim: 48
|
| 123 |
+
dropout: 0.2
|
| 124 |
+
final_dropout: true
|
| 125 |
+
interleave_self_attention: true
|
| 126 |
+
norm_type: ada_norm
|
| 127 |
+
num_attention_heads: 32
|
| 128 |
+
num_layers: 16
|
| 129 |
+
output_dim: 1024
|
| 130 |
+
positional_embeddings: null
|
| 131 |
+
diversity_penalty: 0.0
|
| 132 |
+
do_sample: false
|
| 133 |
+
dtype: null
|
| 134 |
+
early_stopping: false
|
| 135 |
+
encoder_no_repeat_ngram_size: 0
|
| 136 |
+
eos_token_id: null
|
| 137 |
+
exponential_decay_length_penalty: null
|
| 138 |
+
extra_augmentation_config: null
|
| 139 |
+
finetuning_task: null
|
| 140 |
+
forced_bos_token_id: null
|
| 141 |
+
forced_eos_token_id: null
|
| 142 |
+
id2label:
|
| 143 |
+
0: LABEL_0
|
| 144 |
+
1: LABEL_1
|
| 145 |
+
is_decoder: false
|
| 146 |
+
is_encoder_decoder: false
|
| 147 |
+
label2id:
|
| 148 |
+
LABEL_0: 0
|
| 149 |
+
LABEL_1: 1
|
| 150 |
+
length_penalty: 1.0
|
| 151 |
+
load_bf16: false
|
| 152 |
+
max_length: 20
|
| 153 |
+
min_length: 0
|
| 154 |
+
model_name: nvidia/Cosmos-Reason2-2B
|
| 155 |
+
no_repeat_ngram_size: 0
|
| 156 |
+
num_beam_groups: 1
|
| 157 |
+
num_beams: 1
|
| 158 |
+
num_return_sequences: 1
|
| 159 |
+
output_hidden_states: false
|
| 160 |
+
output_scores: false
|
| 161 |
+
pad_token_id: null
|
| 162 |
+
prefix: null
|
| 163 |
+
problem_type: null
|
| 164 |
+
pruned_heads: {}
|
| 165 |
+
random_rotation_angle: null
|
| 166 |
+
remove_invalid_values: false
|
| 167 |
+
repetition_penalty: 1.0
|
| 168 |
+
reproject_vision: false
|
| 169 |
+
return_dict: true
|
| 170 |
+
return_dict_in_generate: false
|
| 171 |
+
sep_token_id: null
|
| 172 |
+
state_dropout_prob: 0.2
|
| 173 |
+
suppress_tokens: null
|
| 174 |
+
task_specific_params: null
|
| 175 |
+
temperature: 1.0
|
| 176 |
+
tf_legacy_loss: false
|
| 177 |
+
tie_encoder_decoder: false
|
| 178 |
+
tie_word_embeddings: true
|
| 179 |
+
tokenizer_class: null
|
| 180 |
+
top_k: 50
|
| 181 |
+
top_p: 1.0
|
| 182 |
+
torchscript: false
|
| 183 |
+
transformers_version: null
|
| 184 |
+
tune_diffusion_model: true
|
| 185 |
+
tune_llm: false
|
| 186 |
+
tune_projector: true
|
| 187 |
+
tune_visual: false
|
| 188 |
+
typical_p: 1.0
|
| 189 |
+
use_bfloat16: false
|
| 190 |
+
use_percentiles: false
|
| 191 |
+
use_relative_action: true
|
| 192 |
+
training: !!python/object:gr00t.configs.training.training_config.TrainingConfig
|
| 193 |
+
add_rl_callback: false
|
| 194 |
+
assert_loss_less_than: null
|
| 195 |
+
batch_size: null
|
| 196 |
+
bf16: true
|
| 197 |
+
dataloader_num_workers: 8
|
| 198 |
+
ddp_bucket_cap_mb: 100
|
| 199 |
+
deepspeed_stage: 2
|
| 200 |
+
enable_open_loop_eval: false
|
| 201 |
+
enable_profiling: false
|
| 202 |
+
eval_batch_size: 2
|
| 203 |
+
eval_bf16: true
|
| 204 |
+
eval_set_split_ratio: 0.1
|
| 205 |
+
eval_steps: 500
|
| 206 |
+
eval_strategy: 'no'
|
| 207 |
+
experiment_name: task0025_subtask_lang_final_layer_vlm_threecam_minmax_80k_gpu1_20260618_045240
|
| 208 |
+
fp16: false
|
| 209 |
+
global_batch_size: 8
|
| 210 |
+
gradient_accumulation_steps: 1
|
| 211 |
+
gradient_checkpointing: false
|
| 212 |
+
learning_rate: 0.0001
|
| 213 |
+
logging_steps: 10
|
| 214 |
+
lr_scheduler_type: cosine
|
| 215 |
+
max_concurrent_uploads: 2
|
| 216 |
+
max_grad_norm: 1.0
|
| 217 |
+
max_retries: 3
|
| 218 |
+
max_steps: 80000
|
| 219 |
+
num_gpus: 1
|
| 220 |
+
open_loop_eval_plot_indices: null
|
| 221 |
+
open_loop_eval_steps_per_traj: 100
|
| 222 |
+
open_loop_eval_traj_ids:
|
| 223 |
+
- 0
|
| 224 |
+
optim: adamw_torch
|
| 225 |
+
output_dir: /data/checkpoints/task0025_subtask_lang_final_layer_vlm_threecam_minmax_80k_gpu1_20260618_045240
|
| 226 |
+
remove_unused_columns: false
|
| 227 |
+
save_best_eval_metric_greater_is_better: true
|
| 228 |
+
save_best_eval_metric_name: ''
|
| 229 |
+
save_only_model: true
|
| 230 |
+
save_steps: 20000
|
| 231 |
+
save_total_limit: 2
|
| 232 |
+
save_vl_model: false
|
| 233 |
+
skip_weight_loading: false
|
| 234 |
+
start_from_checkpoint: /data/base_model
|
| 235 |
+
tf32: true
|
| 236 |
+
transformers_access_token: null
|
| 237 |
+
transformers_cache_dir: null
|
| 238 |
+
transformers_local_files_only: false
|
| 239 |
+
transformers_trust_remote_code: true
|
| 240 |
+
upload_checkpoints: false
|
| 241 |
+
upload_every: 1000
|
| 242 |
+
upload_last_n_checkpoints: 5
|
| 243 |
+
use_ddp: false
|
| 244 |
+
use_wandb: false
|
| 245 |
+
wandb_project: finetune-gr00t-n1d7
|
| 246 |
+
warmup_ratio: 0.05
|
| 247 |
+
warmup_steps: 0
|
| 248 |
+
weight_decay: 1.0e-05
|
*.pth/experiment_cfg/dataset_statistics.json
ADDED
|
@@ -0,0 +1,321 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"new_embodiment": {
|
| 3 |
+
"state": {
|
| 4 |
+
"arm_left": {
|
| 5 |
+
"min": [
|
| 6 |
+
-0.5475472807884216,
|
| 7 |
+
-0.0003355582884978503,
|
| 8 |
+
-1.3053457736968994,
|
| 9 |
+
-2.6803677082061768,
|
| 10 |
+
-1.6582931280136108,
|
| 11 |
+
-0.6229879260063171,
|
| 12 |
+
-1.4020036458969116,
|
| 13 |
+
-0.0015339808305725455
|
| 14 |
+
],
|
| 15 |
+
"max": [
|
| 16 |
+
1.113825798034668,
|
| 17 |
+
0.8176237344741821,
|
| 18 |
+
0.23757527768611908,
|
| 19 |
+
-0.5585607290267944,
|
| 20 |
+
1.0988095998764038,
|
| 21 |
+
1.172776222229004,
|
| 22 |
+
1.3638184070587158,
|
| 23 |
+
1.0921943187713623
|
| 24 |
+
],
|
| 25 |
+
"mean": [
|
| 26 |
+
0.35580819845199585,
|
| 27 |
+
0.5201765298843384,
|
| 28 |
+
-0.1417115330696106,
|
| 29 |
+
-1.9100525379180908,
|
| 30 |
+
0.06096320226788521,
|
| 31 |
+
0.3721380829811096,
|
| 32 |
+
-0.7635477185249329,
|
| 33 |
+
0.2815868854522705
|
| 34 |
+
],
|
| 35 |
+
"std": [
|
| 36 |
+
0.4073781967163086,
|
| 37 |
+
0.23927268385887146,
|
| 38 |
+
0.2344869077205658,
|
| 39 |
+
0.48143312335014343,
|
| 40 |
+
0.5435206294059753,
|
| 41 |
+
0.27136287093162537,
|
| 42 |
+
0.7339828610420227,
|
| 43 |
+
0.41869908571243286
|
| 44 |
+
],
|
| 45 |
+
"q01": [
|
| 46 |
+
-0.3928197732567787,
|
| 47 |
+
0.026710080318152904,
|
| 48 |
+
-0.9970582705736161,
|
| 49 |
+
-2.6524085998535156,
|
| 50 |
+
-1.257974841594696,
|
| 51 |
+
-0.20474184960126876,
|
| 52 |
+
-1.3634459972381592,
|
| 53 |
+
0.0
|
| 54 |
+
],
|
| 55 |
+
"q99": [
|
| 56 |
+
0.9335210216045381,
|
| 57 |
+
0.8175997734069824,
|
| 58 |
+
0.09205574311316017,
|
| 59 |
+
-0.7039075952768324,
|
| 60 |
+
0.8392134261131291,
|
| 61 |
+
0.997027575969696,
|
| 62 |
+
1.1889616501331333,
|
| 63 |
+
1.0477088689804077
|
| 64 |
+
]
|
| 65 |
+
},
|
| 66 |
+
"arm_right": {
|
| 67 |
+
"min": [
|
| 68 |
+
-0.09659285098314285,
|
| 69 |
+
-0.8176716566085815,
|
| 70 |
+
-0.3275049030780792,
|
| 71 |
+
-2.6530556678771973,
|
| 72 |
+
-0.4579651653766632,
|
| 73 |
+
-0.3955034017562866,
|
| 74 |
+
-0.1349928081035614,
|
| 75 |
+
0.0
|
| 76 |
+
],
|
| 77 |
+
"max": [
|
| 78 |
+
0.8757112622261047,
|
| 79 |
+
-0.08124106377363205,
|
| 80 |
+
1.1064674854278564,
|
| 81 |
+
-1.2425124645233154,
|
| 82 |
+
1.333017349243164,
|
| 83 |
+
0.8877794146537781,
|
| 84 |
+
1.3636425733566284,
|
| 85 |
+
0.8682331442832947
|
| 86 |
+
],
|
| 87 |
+
"mean": [
|
| 88 |
+
0.5309473276138306,
|
| 89 |
+
-0.6017705202102661,
|
| 90 |
+
0.07096648961305618,
|
| 91 |
+
-2.2108590602874756,
|
| 92 |
+
0.26930341124534607,
|
| 93 |
+
0.3429528474807739,
|
| 94 |
+
1.2567896842956543,
|
| 95 |
+
0.04772842675447464
|
| 96 |
+
],
|
| 97 |
+
"std": [
|
| 98 |
+
0.3694664239883423,
|
| 99 |
+
0.16356231272220612,
|
| 100 |
+
0.16321000456809998,
|
| 101 |
+
0.4757940173149109,
|
| 102 |
+
0.4936450123786926,
|
| 103 |
+
0.19132843613624573,
|
| 104 |
+
0.2600846588611603,
|
| 105 |
+
0.14340612292289734
|
| 106 |
+
],
|
| 107 |
+
"q01": [
|
| 108 |
+
0.030643662437796593,
|
| 109 |
+
-0.8176237344741821,
|
| 110 |
+
-0.08104931563138962,
|
| 111 |
+
-2.652324676513672,
|
| 112 |
+
-0.37889325618743896,
|
| 113 |
+
-0.11650907695293426,
|
| 114 |
+
0.21011178195476532,
|
| 115 |
+
0.0
|
| 116 |
+
],
|
| 117 |
+
"q99": [
|
| 118 |
+
0.859029233455658,
|
| 119 |
+
-0.16543731629848477,
|
| 120 |
+
0.8615535879135133,
|
| 121 |
+
-1.5723973023891449,
|
| 122 |
+
1.0613687527179718,
|
| 123 |
+
0.6687612211704257,
|
| 124 |
+
1.3634459972381592,
|
| 125 |
+
0.6028544306755066
|
| 126 |
+
]
|
| 127 |
+
},
|
| 128 |
+
"odometry": {
|
| 129 |
+
"min": [
|
| 130 |
+
-0.015251385979354382,
|
| 131 |
+
-0.3666536211967468,
|
| 132 |
+
-0.24855561554431915
|
| 133 |
+
],
|
| 134 |
+
"max": [
|
| 135 |
+
0.28565824031829834,
|
| 136 |
+
0.019846245646476746,
|
| 137 |
+
0.072826087474823
|
| 138 |
+
],
|
| 139 |
+
"mean": [
|
| 140 |
+
4.3003801692975685e-05,
|
| 141 |
+
-0.027965428307652473,
|
| 142 |
+
-7.66340053814929e-06
|
| 143 |
+
],
|
| 144 |
+
"std": [
|
| 145 |
+
0.002784445881843567,
|
| 146 |
+
0.09051357954740524,
|
| 147 |
+
0.002256158972159028
|
| 148 |
+
],
|
| 149 |
+
"q01": [
|
| 150 |
+
-3.177549693646142e-05,
|
| 151 |
+
-0.3368525901436806,
|
| 152 |
+
-0.0052148745534941554
|
| 153 |
+
],
|
| 154 |
+
"q99": [
|
| 155 |
+
3.881613905832974e-05,
|
| 156 |
+
0.0012015580607112498,
|
| 157 |
+
0.005100751668214812
|
| 158 |
+
]
|
| 159 |
+
}
|
| 160 |
+
},
|
| 161 |
+
"action": {
|
| 162 |
+
"arm_left": {
|
| 163 |
+
"min": [
|
| 164 |
+
-0.5476311445236206,
|
| 165 |
+
0.0,
|
| 166 |
+
-1.3054176568984985,
|
| 167 |
+
-2.682774782180786,
|
| 168 |
+
-1.6628351211547852,
|
| 169 |
+
-0.6258641481399536,
|
| 170 |
+
-1.402058482170105,
|
| 171 |
+
0.0
|
| 172 |
+
],
|
| 173 |
+
"max": [
|
| 174 |
+
1.1172550916671753,
|
| 175 |
+
0.8175997734069824,
|
| 176 |
+
0.2411240190267563,
|
| 177 |
+
-0.5568350553512573,
|
| 178 |
+
1.0998642444610596,
|
| 179 |
+
1.1734952926635742,
|
| 180 |
+
1.4005244970321655,
|
| 181 |
+
1.100000023841858
|
| 182 |
+
],
|
| 183 |
+
"mean": [
|
| 184 |
+
0.35810500383377075,
|
| 185 |
+
0.519511342048645,
|
| 186 |
+
-0.14196135103702545,
|
| 187 |
+
-1.9126713275909424,
|
| 188 |
+
0.058261577039957047,
|
| 189 |
+
0.37303486466407776,
|
| 190 |
+
-0.763543426990509,
|
| 191 |
+
0.29336875677108765
|
| 192 |
+
],
|
| 193 |
+
"std": [
|
| 194 |
+
0.40803930163383484,
|
| 195 |
+
0.23888689279556274,
|
| 196 |
+
0.23461289703845978,
|
| 197 |
+
0.4830642342567444,
|
| 198 |
+
0.545210599899292,
|
| 199 |
+
0.2715252637863159,
|
| 200 |
+
0.7340782880783081,
|
| 201 |
+
0.44709646701812744
|
| 202 |
+
],
|
| 203 |
+
"q01": [
|
| 204 |
+
-0.3942330479621887,
|
| 205 |
+
0.026077672839164734,
|
| 206 |
+
-0.9986214637756348,
|
| 207 |
+
-2.652252674102783,
|
| 208 |
+
-1.2593982219696045,
|
| 209 |
+
-0.2070874124765396,
|
| 210 |
+
-1.3634459972381592,
|
| 211 |
+
0.0
|
| 212 |
+
],
|
| 213 |
+
"q99": [
|
| 214 |
+
0.9341943264007568,
|
| 215 |
+
0.8175997734069824,
|
| 216 |
+
0.09348710186779549,
|
| 217 |
+
-0.7034682774543709,
|
| 218 |
+
0.8390874862670898,
|
| 219 |
+
0.9986214637756348,
|
| 220 |
+
1.1903691291809082,
|
| 221 |
+
1.100000023841858
|
| 222 |
+
]
|
| 223 |
+
},
|
| 224 |
+
"arm_right": {
|
| 225 |
+
"min": [
|
| 226 |
+
-0.09980137646198273,
|
| 227 |
+
-0.8175997734069824,
|
| 228 |
+
-0.32994207739830017,
|
| 229 |
+
-2.652252674102783,
|
| 230 |
+
-0.46202123165130615,
|
| 231 |
+
-0.3957670331001282,
|
| 232 |
+
-0.13499030470848083,
|
| 233 |
+
0.0
|
| 234 |
+
],
|
| 235 |
+
"max": [
|
| 236 |
+
0.8774369955062866,
|
| 237 |
+
-0.08130098134279251,
|
| 238 |
+
1.1075341701507568,
|
| 239 |
+
-1.2394565343856812,
|
| 240 |
+
1.3330292701721191,
|
| 241 |
+
0.8881748914718628,
|
| 242 |
+
1.3634459972381592,
|
| 243 |
+
0.8795690536499023
|
| 244 |
+
],
|
| 245 |
+
"mean": [
|
| 246 |
+
0.5329897999763489,
|
| 247 |
+
-0.6011292338371277,
|
| 248 |
+
0.07123955339193344,
|
| 249 |
+
-2.2133889198303223,
|
| 250 |
+
0.27188241481781006,
|
| 251 |
+
0.3438069224357605,
|
| 252 |
+
1.2567932605743408,
|
| 253 |
+
0.04718247056007385
|
| 254 |
+
],
|
| 255 |
+
"std": [
|
| 256 |
+
0.3692541718482971,
|
| 257 |
+
0.1632792353630066,
|
| 258 |
+
0.16333064436912537,
|
| 259 |
+
0.4753437936306,
|
| 260 |
+
0.4933122396469116,
|
| 261 |
+
0.1913645714521408,
|
| 262 |
+
0.2600988447666168,
|
| 263 |
+
0.14372898638248444
|
| 264 |
+
],
|
| 265 |
+
"q01": [
|
| 266 |
+
0.030667630955576897,
|
| 267 |
+
-0.8175997734069824,
|
| 268 |
+
-0.08166213139891625,
|
| 269 |
+
-2.652252674102783,
|
| 270 |
+
-0.3788812756538391,
|
| 271 |
+
-0.11658254265785217,
|
| 272 |
+
0.21015536785125732,
|
| 273 |
+
0.0
|
| 274 |
+
],
|
| 275 |
+
"q99": [
|
| 276 |
+
0.8590052723884583,
|
| 277 |
+
-0.16566991806030273,
|
| 278 |
+
0.8620972037315369,
|
| 279 |
+
-1.5707963705062866,
|
| 280 |
+
1.0630487203598022,
|
| 281 |
+
0.6688156127929688,
|
| 282 |
+
1.3634459972381592,
|
| 283 |
+
0.6062136888504028
|
| 284 |
+
]
|
| 285 |
+
},
|
| 286 |
+
"odometry": {
|
| 287 |
+
"min": [
|
| 288 |
+
0.0,
|
| 289 |
+
-0.330078125,
|
| 290 |
+
-0.4609375
|
| 291 |
+
],
|
| 292 |
+
"max": [
|
| 293 |
+
0.3072916567325592,
|
| 294 |
+
-0.0,
|
| 295 |
+
-0.0
|
| 296 |
+
],
|
| 297 |
+
"mean": [
|
| 298 |
+
6.987077358644456e-05,
|
| 299 |
+
-0.028174595907330513,
|
| 300 |
+
-9.38360062718857e-06
|
| 301 |
+
],
|
| 302 |
+
"std": [
|
| 303 |
+
0.004499922972172499,
|
| 304 |
+
0.09222032129764557,
|
| 305 |
+
0.002032066462561488
|
| 306 |
+
],
|
| 307 |
+
"q01": [
|
| 308 |
+
0.0,
|
| 309 |
+
-0.330078125,
|
| 310 |
+
-0.0
|
| 311 |
+
],
|
| 312 |
+
"q99": [
|
| 313 |
+
0.0,
|
| 314 |
+
0.0,
|
| 315 |
+
0.0
|
| 316 |
+
]
|
| 317 |
+
}
|
| 318 |
+
},
|
| 319 |
+
"relative_action": {}
|
| 320 |
+
}
|
| 321 |
+
}
|
*.pth/experiment_cfg/final_model_config.json
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"model_type": "Gr00tN1d7",
|
| 3 |
+
"model_dtype": "bfloat16",
|
| 4 |
+
"model_name": "nvidia/Cosmos-Reason2-2B",
|
| 5 |
+
"backbone_model_type": "qwen",
|
| 6 |
+
"model_revision": null,
|
| 7 |
+
"tune_top_llm_layers": 0,
|
| 8 |
+
"backbone_embedding_dim": 2048,
|
| 9 |
+
"tune_llm": false,
|
| 10 |
+
"tune_visual": false,
|
| 11 |
+
"select_layer": 16,
|
| 12 |
+
"reproject_vision": false,
|
| 13 |
+
"use_flash_attention": true,
|
| 14 |
+
"load_bf16": false,
|
| 15 |
+
"backbone_trainable_params_fp32": true,
|
| 16 |
+
"extra_augmentation_config": null,
|
| 17 |
+
"apply_sincos_state_encoding": false,
|
| 18 |
+
"use_percentiles": false,
|
| 19 |
+
"clip_outliers": true,
|
| 20 |
+
"use_relative_action": true,
|
| 21 |
+
"max_state_dim": 132,
|
| 22 |
+
"max_action_dim": 132,
|
| 23 |
+
"action_horizon": 40,
|
| 24 |
+
"hidden_size": 1024,
|
| 25 |
+
"input_embedding_dim": 1536,
|
| 26 |
+
"state_history_length": 1,
|
| 27 |
+
"add_pos_embed": true,
|
| 28 |
+
"attn_dropout": 0.2,
|
| 29 |
+
"use_vlln": true,
|
| 30 |
+
"max_seq_len": 1024,
|
| 31 |
+
"use_alternate_vl_dit": true,
|
| 32 |
+
"attend_text_every_n_blocks": 2,
|
| 33 |
+
"diffusion_model_cfg": {
|
| 34 |
+
"attention_head_dim": 48,
|
| 35 |
+
"dropout": 0.2,
|
| 36 |
+
"final_dropout": true,
|
| 37 |
+
"interleave_self_attention": true,
|
| 38 |
+
"norm_type": "ada_norm",
|
| 39 |
+
"num_attention_heads": 32,
|
| 40 |
+
"num_layers": 32,
|
| 41 |
+
"output_dim": 1024,
|
| 42 |
+
"positional_embeddings": null
|
| 43 |
+
},
|
| 44 |
+
"num_inference_timesteps": 4,
|
| 45 |
+
"noise_beta_alpha": 1.5,
|
| 46 |
+
"noise_beta_beta": 1.0,
|
| 47 |
+
"noise_s": 0.999,
|
| 48 |
+
"num_timestep_buckets": 1000,
|
| 49 |
+
"tune_projector": true,
|
| 50 |
+
"tune_diffusion_model": true,
|
| 51 |
+
"tune_vlln": true,
|
| 52 |
+
"state_dropout_prob": 0.2,
|
| 53 |
+
"exclude_state": false,
|
| 54 |
+
"use_mean_std": false,
|
| 55 |
+
"max_num_embodiments": 32
|
| 56 |
+
}
|
*.pth/experiment_cfg/final_processor_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
*.pth/model-00001-of-00003.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1f2fe11062392310da5ef217d8143d9f870f0547c0e0985b54a87c5af861d184
|
| 3 |
+
size 4986649584
|
*.pth/model-00002-of-00003.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5e509688d6023afda10e33c3f99919e24c04a2f53325425ff4d013734e84476c
|
| 3 |
+
size 4970792616
|
*.pth/model-00003-of-00003.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:837e279fac4c925b3878b80bebfdc68337022cc487b8ac611cc10cf0a63fdc1c
|
| 3 |
+
size 2618758696
|
*.pth/model.safetensors.index.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
*.pth/processor_config.json
ADDED
|
@@ -0,0 +1,1132 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"processor_class": "Gr00tN1d7Processor",
|
| 3 |
+
"processor_kwargs": {
|
| 4 |
+
"modality_configs": {
|
| 5 |
+
"real_g1_relative_eef_relative_joints": {
|
| 6 |
+
"video": {
|
| 7 |
+
"delta_indices": [
|
| 8 |
+
-20,
|
| 9 |
+
0
|
| 10 |
+
],
|
| 11 |
+
"modality_keys": [
|
| 12 |
+
"ego_view"
|
| 13 |
+
],
|
| 14 |
+
"sin_cos_embedding_keys": null,
|
| 15 |
+
"mean_std_embedding_keys": null,
|
| 16 |
+
"action_configs": null
|
| 17 |
+
},
|
| 18 |
+
"state": {
|
| 19 |
+
"delta_indices": [
|
| 20 |
+
0
|
| 21 |
+
],
|
| 22 |
+
"modality_keys": [
|
| 23 |
+
"left_wrist_eef_9d",
|
| 24 |
+
"right_wrist_eef_9d",
|
| 25 |
+
"left_hand",
|
| 26 |
+
"right_hand",
|
| 27 |
+
"left_arm",
|
| 28 |
+
"right_arm",
|
| 29 |
+
"waist"
|
| 30 |
+
],
|
| 31 |
+
"sin_cos_embedding_keys": null,
|
| 32 |
+
"mean_std_embedding_keys": null,
|
| 33 |
+
"action_configs": null
|
| 34 |
+
},
|
| 35 |
+
"action": {
|
| 36 |
+
"delta_indices": [
|
| 37 |
+
0,
|
| 38 |
+
1,
|
| 39 |
+
2,
|
| 40 |
+
3,
|
| 41 |
+
4,
|
| 42 |
+
5,
|
| 43 |
+
6,
|
| 44 |
+
7,
|
| 45 |
+
8,
|
| 46 |
+
9,
|
| 47 |
+
10,
|
| 48 |
+
11,
|
| 49 |
+
12,
|
| 50 |
+
13,
|
| 51 |
+
14,
|
| 52 |
+
15,
|
| 53 |
+
16,
|
| 54 |
+
17,
|
| 55 |
+
18,
|
| 56 |
+
19,
|
| 57 |
+
20,
|
| 58 |
+
21,
|
| 59 |
+
22,
|
| 60 |
+
23,
|
| 61 |
+
24,
|
| 62 |
+
25,
|
| 63 |
+
26,
|
| 64 |
+
27,
|
| 65 |
+
28,
|
| 66 |
+
29,
|
| 67 |
+
30,
|
| 68 |
+
31,
|
| 69 |
+
32,
|
| 70 |
+
33,
|
| 71 |
+
34,
|
| 72 |
+
35,
|
| 73 |
+
36,
|
| 74 |
+
37,
|
| 75 |
+
38,
|
| 76 |
+
39
|
| 77 |
+
],
|
| 78 |
+
"modality_keys": [
|
| 79 |
+
"left_wrist_eef_9d",
|
| 80 |
+
"right_wrist_eef_9d",
|
| 81 |
+
"left_hand",
|
| 82 |
+
"right_hand",
|
| 83 |
+
"left_arm",
|
| 84 |
+
"right_arm",
|
| 85 |
+
"waist",
|
| 86 |
+
"base_height_command",
|
| 87 |
+
"navigate_command"
|
| 88 |
+
],
|
| 89 |
+
"sin_cos_embedding_keys": null,
|
| 90 |
+
"mean_std_embedding_keys": null,
|
| 91 |
+
"action_configs": [
|
| 92 |
+
{
|
| 93 |
+
"rep": "RELATIVE",
|
| 94 |
+
"type": "EEF",
|
| 95 |
+
"format": "XYZ_ROT6D",
|
| 96 |
+
"state_key": "left_wrist_eef_9d"
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"rep": "RELATIVE",
|
| 100 |
+
"type": "EEF",
|
| 101 |
+
"format": "XYZ_ROT6D",
|
| 102 |
+
"state_key": "right_wrist_eef_9d"
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"rep": "ABSOLUTE",
|
| 106 |
+
"type": "NON_EEF",
|
| 107 |
+
"format": "DEFAULT",
|
| 108 |
+
"state_key": "left_hand"
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"rep": "ABSOLUTE",
|
| 112 |
+
"type": "NON_EEF",
|
| 113 |
+
"format": "DEFAULT",
|
| 114 |
+
"state_key": "right_hand"
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"rep": "RELATIVE",
|
| 118 |
+
"type": "NON_EEF",
|
| 119 |
+
"format": "DEFAULT",
|
| 120 |
+
"state_key": "left_arm"
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"rep": "RELATIVE",
|
| 124 |
+
"type": "NON_EEF",
|
| 125 |
+
"format": "DEFAULT",
|
| 126 |
+
"state_key": "right_arm"
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"rep": "ABSOLUTE",
|
| 130 |
+
"type": "NON_EEF",
|
| 131 |
+
"format": "DEFAULT",
|
| 132 |
+
"state_key": "waist"
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"rep": "ABSOLUTE",
|
| 136 |
+
"type": "NON_EEF",
|
| 137 |
+
"format": "DEFAULT",
|
| 138 |
+
"state_key": "base_height_command"
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"rep": "ABSOLUTE",
|
| 142 |
+
"type": "NON_EEF",
|
| 143 |
+
"format": "DEFAULT",
|
| 144 |
+
"state_key": "navigate_command"
|
| 145 |
+
}
|
| 146 |
+
]
|
| 147 |
+
},
|
| 148 |
+
"language": {
|
| 149 |
+
"delta_indices": [
|
| 150 |
+
0
|
| 151 |
+
],
|
| 152 |
+
"modality_keys": [
|
| 153 |
+
"annotation.human.task_description"
|
| 154 |
+
],
|
| 155 |
+
"sin_cos_embedding_keys": null,
|
| 156 |
+
"mean_std_embedding_keys": null,
|
| 157 |
+
"action_configs": null
|
| 158 |
+
}
|
| 159 |
+
},
|
| 160 |
+
"real_r1_pro_sharpa_relative_eef_mecka": {
|
| 161 |
+
"video": {
|
| 162 |
+
"delta_indices": [
|
| 163 |
+
-30,
|
| 164 |
+
0
|
| 165 |
+
],
|
| 166 |
+
"modality_keys": [
|
| 167 |
+
"ego_view_cropratio_res320x240_freq30"
|
| 168 |
+
],
|
| 169 |
+
"sin_cos_embedding_keys": null,
|
| 170 |
+
"mean_std_embedding_keys": null,
|
| 171 |
+
"action_configs": null
|
| 172 |
+
},
|
| 173 |
+
"state": {
|
| 174 |
+
"delta_indices": [
|
| 175 |
+
0
|
| 176 |
+
],
|
| 177 |
+
"modality_keys": [
|
| 178 |
+
"left_wrist_eef",
|
| 179 |
+
"right_wrist_eef",
|
| 180 |
+
"left_hand_joints",
|
| 181 |
+
"right_hand_joints"
|
| 182 |
+
],
|
| 183 |
+
"sin_cos_embedding_keys": null,
|
| 184 |
+
"mean_std_embedding_keys": null,
|
| 185 |
+
"action_configs": null
|
| 186 |
+
},
|
| 187 |
+
"action": {
|
| 188 |
+
"delta_indices": [
|
| 189 |
+
0,
|
| 190 |
+
1,
|
| 191 |
+
2,
|
| 192 |
+
3,
|
| 193 |
+
4,
|
| 194 |
+
5,
|
| 195 |
+
6,
|
| 196 |
+
7,
|
| 197 |
+
8,
|
| 198 |
+
9,
|
| 199 |
+
10,
|
| 200 |
+
11,
|
| 201 |
+
12,
|
| 202 |
+
13,
|
| 203 |
+
14,
|
| 204 |
+
15,
|
| 205 |
+
16,
|
| 206 |
+
17,
|
| 207 |
+
18,
|
| 208 |
+
19,
|
| 209 |
+
20,
|
| 210 |
+
21,
|
| 211 |
+
22,
|
| 212 |
+
23,
|
| 213 |
+
24,
|
| 214 |
+
25,
|
| 215 |
+
26,
|
| 216 |
+
27,
|
| 217 |
+
28,
|
| 218 |
+
29,
|
| 219 |
+
30,
|
| 220 |
+
31,
|
| 221 |
+
32,
|
| 222 |
+
33,
|
| 223 |
+
34,
|
| 224 |
+
35,
|
| 225 |
+
36,
|
| 226 |
+
37,
|
| 227 |
+
38,
|
| 228 |
+
39
|
| 229 |
+
],
|
| 230 |
+
"modality_keys": [
|
| 231 |
+
"left_wrist_eef",
|
| 232 |
+
"right_wrist_eef",
|
| 233 |
+
"left_hand_joints",
|
| 234 |
+
"right_hand_joints"
|
| 235 |
+
],
|
| 236 |
+
"sin_cos_embedding_keys": null,
|
| 237 |
+
"mean_std_embedding_keys": null,
|
| 238 |
+
"action_configs": [
|
| 239 |
+
{
|
| 240 |
+
"rep": "RELATIVE",
|
| 241 |
+
"type": "EEF",
|
| 242 |
+
"format": "XYZ_ROT6D",
|
| 243 |
+
"state_key": "left_wrist_eef"
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"rep": "RELATIVE",
|
| 247 |
+
"type": "EEF",
|
| 248 |
+
"format": "XYZ_ROT6D",
|
| 249 |
+
"state_key": "right_wrist_eef"
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"rep": "ABSOLUTE",
|
| 253 |
+
"type": "NON_EEF",
|
| 254 |
+
"format": "DEFAULT",
|
| 255 |
+
"state_key": "left_hand_joints"
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"rep": "ABSOLUTE",
|
| 259 |
+
"type": "NON_EEF",
|
| 260 |
+
"format": "DEFAULT",
|
| 261 |
+
"state_key": "right_hand_joints"
|
| 262 |
+
}
|
| 263 |
+
]
|
| 264 |
+
},
|
| 265 |
+
"language": {
|
| 266 |
+
"delta_indices": [
|
| 267 |
+
0
|
| 268 |
+
],
|
| 269 |
+
"modality_keys": [
|
| 270 |
+
"annotation.human.coarse_action"
|
| 271 |
+
],
|
| 272 |
+
"sin_cos_embedding_keys": null,
|
| 273 |
+
"mean_std_embedding_keys": null,
|
| 274 |
+
"action_configs": null
|
| 275 |
+
}
|
| 276 |
+
},
|
| 277 |
+
"real_r1_pro_sharpa_relative_eef_human": {
|
| 278 |
+
"video": {
|
| 279 |
+
"delta_indices": [
|
| 280 |
+
-20,
|
| 281 |
+
0
|
| 282 |
+
],
|
| 283 |
+
"modality_keys": [
|
| 284 |
+
"ego_view_res320x240_freq20",
|
| 285 |
+
"left_wrist_view_res320x240_freq20",
|
| 286 |
+
"right_wrist_view_res320x240_freq20"
|
| 287 |
+
],
|
| 288 |
+
"sin_cos_embedding_keys": null,
|
| 289 |
+
"mean_std_embedding_keys": null,
|
| 290 |
+
"action_configs": null
|
| 291 |
+
},
|
| 292 |
+
"state": {
|
| 293 |
+
"delta_indices": [
|
| 294 |
+
0
|
| 295 |
+
],
|
| 296 |
+
"modality_keys": [
|
| 297 |
+
"left_wrist_eef",
|
| 298 |
+
"right_wrist_eef",
|
| 299 |
+
"left_hand_joints",
|
| 300 |
+
"right_hand_joints"
|
| 301 |
+
],
|
| 302 |
+
"sin_cos_embedding_keys": null,
|
| 303 |
+
"mean_std_embedding_keys": null,
|
| 304 |
+
"action_configs": null
|
| 305 |
+
},
|
| 306 |
+
"action": {
|
| 307 |
+
"delta_indices": [
|
| 308 |
+
0,
|
| 309 |
+
1,
|
| 310 |
+
2,
|
| 311 |
+
3,
|
| 312 |
+
4,
|
| 313 |
+
5,
|
| 314 |
+
6,
|
| 315 |
+
7,
|
| 316 |
+
8,
|
| 317 |
+
9,
|
| 318 |
+
10,
|
| 319 |
+
11,
|
| 320 |
+
12,
|
| 321 |
+
13,
|
| 322 |
+
14,
|
| 323 |
+
15,
|
| 324 |
+
16,
|
| 325 |
+
17,
|
| 326 |
+
18,
|
| 327 |
+
19,
|
| 328 |
+
20,
|
| 329 |
+
21,
|
| 330 |
+
22,
|
| 331 |
+
23,
|
| 332 |
+
24,
|
| 333 |
+
25,
|
| 334 |
+
26,
|
| 335 |
+
27,
|
| 336 |
+
28,
|
| 337 |
+
29,
|
| 338 |
+
30,
|
| 339 |
+
31,
|
| 340 |
+
32,
|
| 341 |
+
33,
|
| 342 |
+
34,
|
| 343 |
+
35,
|
| 344 |
+
36,
|
| 345 |
+
37,
|
| 346 |
+
38,
|
| 347 |
+
39
|
| 348 |
+
],
|
| 349 |
+
"modality_keys": [
|
| 350 |
+
"left_wrist_eef",
|
| 351 |
+
"right_wrist_eef",
|
| 352 |
+
"left_hand_joints",
|
| 353 |
+
"right_hand_joints"
|
| 354 |
+
],
|
| 355 |
+
"sin_cos_embedding_keys": null,
|
| 356 |
+
"mean_std_embedding_keys": null,
|
| 357 |
+
"action_configs": [
|
| 358 |
+
{
|
| 359 |
+
"rep": "RELATIVE",
|
| 360 |
+
"type": "EEF",
|
| 361 |
+
"format": "XYZ_ROT6D",
|
| 362 |
+
"state_key": "left_wrist_eef"
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"rep": "RELATIVE",
|
| 366 |
+
"type": "EEF",
|
| 367 |
+
"format": "XYZ_ROT6D",
|
| 368 |
+
"state_key": "right_wrist_eef"
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"rep": "ABSOLUTE",
|
| 372 |
+
"type": "NON_EEF",
|
| 373 |
+
"format": "DEFAULT",
|
| 374 |
+
"state_key": "left_hand_joints"
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"rep": "ABSOLUTE",
|
| 378 |
+
"type": "NON_EEF",
|
| 379 |
+
"format": "DEFAULT",
|
| 380 |
+
"state_key": "right_hand_joints"
|
| 381 |
+
}
|
| 382 |
+
]
|
| 383 |
+
},
|
| 384 |
+
"language": {
|
| 385 |
+
"delta_indices": [
|
| 386 |
+
0
|
| 387 |
+
],
|
| 388 |
+
"modality_keys": [
|
| 389 |
+
"annotation.human.coarse_action"
|
| 390 |
+
],
|
| 391 |
+
"sin_cos_embedding_keys": null,
|
| 392 |
+
"mean_std_embedding_keys": null,
|
| 393 |
+
"action_configs": null
|
| 394 |
+
}
|
| 395 |
+
},
|
| 396 |
+
"real_r1_pro_sharpa_relative_eef": {
|
| 397 |
+
"video": {
|
| 398 |
+
"delta_indices": [
|
| 399 |
+
-20,
|
| 400 |
+
0
|
| 401 |
+
],
|
| 402 |
+
"modality_keys": [
|
| 403 |
+
"ego_view_res320x240_freq20",
|
| 404 |
+
"left_wrist_view_res320x240_freq20",
|
| 405 |
+
"right_wrist_view_res320x240_freq20"
|
| 406 |
+
],
|
| 407 |
+
"sin_cos_embedding_keys": null,
|
| 408 |
+
"mean_std_embedding_keys": null,
|
| 409 |
+
"action_configs": null
|
| 410 |
+
},
|
| 411 |
+
"state": {
|
| 412 |
+
"delta_indices": [
|
| 413 |
+
0
|
| 414 |
+
],
|
| 415 |
+
"modality_keys": [
|
| 416 |
+
"left_wrist_eef",
|
| 417 |
+
"right_wrist_eef",
|
| 418 |
+
"left_hand_joints",
|
| 419 |
+
"right_hand_joints"
|
| 420 |
+
],
|
| 421 |
+
"sin_cos_embedding_keys": null,
|
| 422 |
+
"mean_std_embedding_keys": null,
|
| 423 |
+
"action_configs": null
|
| 424 |
+
},
|
| 425 |
+
"action": {
|
| 426 |
+
"delta_indices": [
|
| 427 |
+
0,
|
| 428 |
+
1,
|
| 429 |
+
2,
|
| 430 |
+
3,
|
| 431 |
+
4,
|
| 432 |
+
5,
|
| 433 |
+
6,
|
| 434 |
+
7,
|
| 435 |
+
8,
|
| 436 |
+
9,
|
| 437 |
+
10,
|
| 438 |
+
11,
|
| 439 |
+
12,
|
| 440 |
+
13,
|
| 441 |
+
14,
|
| 442 |
+
15,
|
| 443 |
+
16,
|
| 444 |
+
17,
|
| 445 |
+
18,
|
| 446 |
+
19,
|
| 447 |
+
20,
|
| 448 |
+
21,
|
| 449 |
+
22,
|
| 450 |
+
23,
|
| 451 |
+
24,
|
| 452 |
+
25,
|
| 453 |
+
26,
|
| 454 |
+
27,
|
| 455 |
+
28,
|
| 456 |
+
29,
|
| 457 |
+
30,
|
| 458 |
+
31,
|
| 459 |
+
32,
|
| 460 |
+
33,
|
| 461 |
+
34,
|
| 462 |
+
35,
|
| 463 |
+
36,
|
| 464 |
+
37,
|
| 465 |
+
38,
|
| 466 |
+
39
|
| 467 |
+
],
|
| 468 |
+
"modality_keys": [
|
| 469 |
+
"left_wrist_eef",
|
| 470 |
+
"right_wrist_eef",
|
| 471 |
+
"left_hand_joints",
|
| 472 |
+
"right_hand_joints"
|
| 473 |
+
],
|
| 474 |
+
"sin_cos_embedding_keys": null,
|
| 475 |
+
"mean_std_embedding_keys": null,
|
| 476 |
+
"action_configs": [
|
| 477 |
+
{
|
| 478 |
+
"rep": "RELATIVE",
|
| 479 |
+
"type": "EEF",
|
| 480 |
+
"format": "XYZ_ROT6D",
|
| 481 |
+
"state_key": "left_wrist_eef"
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"rep": "RELATIVE",
|
| 485 |
+
"type": "EEF",
|
| 486 |
+
"format": "XYZ_ROT6D",
|
| 487 |
+
"state_key": "right_wrist_eef"
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"rep": "ABSOLUTE",
|
| 491 |
+
"type": "NON_EEF",
|
| 492 |
+
"format": "DEFAULT",
|
| 493 |
+
"state_key": "left_hand_joints"
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"rep": "ABSOLUTE",
|
| 497 |
+
"type": "NON_EEF",
|
| 498 |
+
"format": "DEFAULT",
|
| 499 |
+
"state_key": "right_hand_joints"
|
| 500 |
+
}
|
| 501 |
+
]
|
| 502 |
+
},
|
| 503 |
+
"language": {
|
| 504 |
+
"delta_indices": [
|
| 505 |
+
0
|
| 506 |
+
],
|
| 507 |
+
"modality_keys": [
|
| 508 |
+
"annotation.human.coarse_action"
|
| 509 |
+
],
|
| 510 |
+
"sin_cos_embedding_keys": null,
|
| 511 |
+
"mean_std_embedding_keys": null,
|
| 512 |
+
"action_configs": null
|
| 513 |
+
}
|
| 514 |
+
},
|
| 515 |
+
"xdof_relative_eef_relative_joint": {
|
| 516 |
+
"video": {
|
| 517 |
+
"delta_indices": [
|
| 518 |
+
-30,
|
| 519 |
+
0
|
| 520 |
+
],
|
| 521 |
+
"modality_keys": [
|
| 522 |
+
"top_camera-images-rgb_320_240",
|
| 523 |
+
"left_camera-images-rgb_320_240",
|
| 524 |
+
"right_camera-images-rgb_320_240"
|
| 525 |
+
],
|
| 526 |
+
"sin_cos_embedding_keys": null,
|
| 527 |
+
"mean_std_embedding_keys": null,
|
| 528 |
+
"action_configs": null
|
| 529 |
+
},
|
| 530 |
+
"state": {
|
| 531 |
+
"delta_indices": [
|
| 532 |
+
0
|
| 533 |
+
],
|
| 534 |
+
"modality_keys": [
|
| 535 |
+
"left_wrist_eef",
|
| 536 |
+
"right_wrist_eef",
|
| 537 |
+
"left_gripper_pos",
|
| 538 |
+
"right_gripper_pos",
|
| 539 |
+
"left_joint_pos",
|
| 540 |
+
"right_joint_pos"
|
| 541 |
+
],
|
| 542 |
+
"sin_cos_embedding_keys": null,
|
| 543 |
+
"mean_std_embedding_keys": null,
|
| 544 |
+
"action_configs": null
|
| 545 |
+
},
|
| 546 |
+
"action": {
|
| 547 |
+
"delta_indices": [
|
| 548 |
+
0,
|
| 549 |
+
1,
|
| 550 |
+
2,
|
| 551 |
+
3,
|
| 552 |
+
4,
|
| 553 |
+
5,
|
| 554 |
+
6,
|
| 555 |
+
7,
|
| 556 |
+
8,
|
| 557 |
+
9,
|
| 558 |
+
10,
|
| 559 |
+
11,
|
| 560 |
+
12,
|
| 561 |
+
13,
|
| 562 |
+
14,
|
| 563 |
+
15,
|
| 564 |
+
16,
|
| 565 |
+
17,
|
| 566 |
+
18,
|
| 567 |
+
19,
|
| 568 |
+
20,
|
| 569 |
+
21,
|
| 570 |
+
22,
|
| 571 |
+
23,
|
| 572 |
+
24,
|
| 573 |
+
25,
|
| 574 |
+
26,
|
| 575 |
+
27,
|
| 576 |
+
28,
|
| 577 |
+
29,
|
| 578 |
+
30,
|
| 579 |
+
31,
|
| 580 |
+
32,
|
| 581 |
+
33,
|
| 582 |
+
34,
|
| 583 |
+
35,
|
| 584 |
+
36,
|
| 585 |
+
37,
|
| 586 |
+
38,
|
| 587 |
+
39
|
| 588 |
+
],
|
| 589 |
+
"modality_keys": [
|
| 590 |
+
"left_wrist_eef",
|
| 591 |
+
"right_wrist_eef",
|
| 592 |
+
"left_gripper_pos",
|
| 593 |
+
"right_gripper_pos",
|
| 594 |
+
"left_joint_pos",
|
| 595 |
+
"right_joint_pos"
|
| 596 |
+
],
|
| 597 |
+
"sin_cos_embedding_keys": null,
|
| 598 |
+
"mean_std_embedding_keys": null,
|
| 599 |
+
"action_configs": [
|
| 600 |
+
{
|
| 601 |
+
"rep": "RELATIVE",
|
| 602 |
+
"type": "EEF",
|
| 603 |
+
"format": "XYZ_ROT6D",
|
| 604 |
+
"state_key": "left_wrist_eef"
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"rep": "RELATIVE",
|
| 608 |
+
"type": "EEF",
|
| 609 |
+
"format": "XYZ_ROT6D",
|
| 610 |
+
"state_key": "right_wrist_eef"
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"rep": "ABSOLUTE",
|
| 614 |
+
"type": "NON_EEF",
|
| 615 |
+
"format": "DEFAULT",
|
| 616 |
+
"state_key": "left_gripper_pos"
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"rep": "ABSOLUTE",
|
| 620 |
+
"type": "NON_EEF",
|
| 621 |
+
"format": "DEFAULT",
|
| 622 |
+
"state_key": "right_gripper_pos"
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"rep": "RELATIVE",
|
| 626 |
+
"type": "NON_EEF",
|
| 627 |
+
"format": "DEFAULT",
|
| 628 |
+
"state_key": "left_joint_pos"
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"rep": "RELATIVE",
|
| 632 |
+
"type": "NON_EEF",
|
| 633 |
+
"format": "DEFAULT",
|
| 634 |
+
"state_key": "right_joint_pos"
|
| 635 |
+
}
|
| 636 |
+
]
|
| 637 |
+
},
|
| 638 |
+
"language": {
|
| 639 |
+
"delta_indices": [
|
| 640 |
+
0
|
| 641 |
+
],
|
| 642 |
+
"modality_keys": [
|
| 643 |
+
"annotation.task"
|
| 644 |
+
],
|
| 645 |
+
"sin_cos_embedding_keys": null,
|
| 646 |
+
"mean_std_embedding_keys": null,
|
| 647 |
+
"action_configs": null
|
| 648 |
+
}
|
| 649 |
+
},
|
| 650 |
+
"real_r1_pro_sharpa_relative_eef_maxinsights": {
|
| 651 |
+
"video": {
|
| 652 |
+
"delta_indices": [
|
| 653 |
+
-30,
|
| 654 |
+
0
|
| 655 |
+
],
|
| 656 |
+
"modality_keys": [
|
| 657 |
+
"ego_view_cropratio_res320x240_freq30"
|
| 658 |
+
],
|
| 659 |
+
"sin_cos_embedding_keys": null,
|
| 660 |
+
"mean_std_embedding_keys": null,
|
| 661 |
+
"action_configs": null
|
| 662 |
+
},
|
| 663 |
+
"state": {
|
| 664 |
+
"delta_indices": [
|
| 665 |
+
0
|
| 666 |
+
],
|
| 667 |
+
"modality_keys": [
|
| 668 |
+
"left_wrist_eef",
|
| 669 |
+
"right_wrist_eef",
|
| 670 |
+
"left_hand_joints",
|
| 671 |
+
"right_hand_joints"
|
| 672 |
+
],
|
| 673 |
+
"sin_cos_embedding_keys": null,
|
| 674 |
+
"mean_std_embedding_keys": null,
|
| 675 |
+
"action_configs": null
|
| 676 |
+
},
|
| 677 |
+
"action": {
|
| 678 |
+
"delta_indices": [
|
| 679 |
+
0,
|
| 680 |
+
1,
|
| 681 |
+
2,
|
| 682 |
+
3,
|
| 683 |
+
4,
|
| 684 |
+
5,
|
| 685 |
+
6,
|
| 686 |
+
7,
|
| 687 |
+
8,
|
| 688 |
+
9,
|
| 689 |
+
10,
|
| 690 |
+
11,
|
| 691 |
+
12,
|
| 692 |
+
13,
|
| 693 |
+
14,
|
| 694 |
+
15,
|
| 695 |
+
16,
|
| 696 |
+
17,
|
| 697 |
+
18,
|
| 698 |
+
19,
|
| 699 |
+
20,
|
| 700 |
+
21,
|
| 701 |
+
22,
|
| 702 |
+
23,
|
| 703 |
+
24,
|
| 704 |
+
25,
|
| 705 |
+
26,
|
| 706 |
+
27,
|
| 707 |
+
28,
|
| 708 |
+
29,
|
| 709 |
+
30,
|
| 710 |
+
31,
|
| 711 |
+
32,
|
| 712 |
+
33,
|
| 713 |
+
34,
|
| 714 |
+
35,
|
| 715 |
+
36,
|
| 716 |
+
37,
|
| 717 |
+
38,
|
| 718 |
+
39
|
| 719 |
+
],
|
| 720 |
+
"modality_keys": [
|
| 721 |
+
"left_wrist_eef",
|
| 722 |
+
"right_wrist_eef",
|
| 723 |
+
"left_hand_joints",
|
| 724 |
+
"right_hand_joints"
|
| 725 |
+
],
|
| 726 |
+
"sin_cos_embedding_keys": null,
|
| 727 |
+
"mean_std_embedding_keys": null,
|
| 728 |
+
"action_configs": [
|
| 729 |
+
{
|
| 730 |
+
"rep": "RELATIVE",
|
| 731 |
+
"type": "EEF",
|
| 732 |
+
"format": "XYZ_ROT6D",
|
| 733 |
+
"state_key": "left_wrist_eef"
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"rep": "RELATIVE",
|
| 737 |
+
"type": "EEF",
|
| 738 |
+
"format": "XYZ_ROT6D",
|
| 739 |
+
"state_key": "right_wrist_eef"
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"rep": "ABSOLUTE",
|
| 743 |
+
"type": "NON_EEF",
|
| 744 |
+
"format": "DEFAULT",
|
| 745 |
+
"state_key": "left_hand_joints"
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"rep": "ABSOLUTE",
|
| 749 |
+
"type": "NON_EEF",
|
| 750 |
+
"format": "DEFAULT",
|
| 751 |
+
"state_key": "right_hand_joints"
|
| 752 |
+
}
|
| 753 |
+
]
|
| 754 |
+
},
|
| 755 |
+
"language": {
|
| 756 |
+
"delta_indices": [
|
| 757 |
+
0
|
| 758 |
+
],
|
| 759 |
+
"modality_keys": [
|
| 760 |
+
"annotation.human.coarse_action"
|
| 761 |
+
],
|
| 762 |
+
"sin_cos_embedding_keys": null,
|
| 763 |
+
"mean_std_embedding_keys": null,
|
| 764 |
+
"action_configs": null
|
| 765 |
+
}
|
| 766 |
+
},
|
| 767 |
+
"xdof_relative_eef_relative_joint_subtask": {
|
| 768 |
+
"video": {
|
| 769 |
+
"delta_indices": [
|
| 770 |
+
-30,
|
| 771 |
+
0
|
| 772 |
+
],
|
| 773 |
+
"modality_keys": [
|
| 774 |
+
"top_camera-images-rgb_320_240",
|
| 775 |
+
"left_camera-images-rgb_320_240",
|
| 776 |
+
"right_camera-images-rgb_320_240"
|
| 777 |
+
],
|
| 778 |
+
"sin_cos_embedding_keys": null,
|
| 779 |
+
"mean_std_embedding_keys": null,
|
| 780 |
+
"action_configs": null
|
| 781 |
+
},
|
| 782 |
+
"state": {
|
| 783 |
+
"delta_indices": [
|
| 784 |
+
0
|
| 785 |
+
],
|
| 786 |
+
"modality_keys": [
|
| 787 |
+
"left_wrist_eef",
|
| 788 |
+
"right_wrist_eef",
|
| 789 |
+
"left_gripper_pos",
|
| 790 |
+
"right_gripper_pos",
|
| 791 |
+
"left_joint_pos",
|
| 792 |
+
"right_joint_pos"
|
| 793 |
+
],
|
| 794 |
+
"sin_cos_embedding_keys": null,
|
| 795 |
+
"mean_std_embedding_keys": null,
|
| 796 |
+
"action_configs": null
|
| 797 |
+
},
|
| 798 |
+
"action": {
|
| 799 |
+
"delta_indices": [
|
| 800 |
+
0,
|
| 801 |
+
1,
|
| 802 |
+
2,
|
| 803 |
+
3,
|
| 804 |
+
4,
|
| 805 |
+
5,
|
| 806 |
+
6,
|
| 807 |
+
7,
|
| 808 |
+
8,
|
| 809 |
+
9,
|
| 810 |
+
10,
|
| 811 |
+
11,
|
| 812 |
+
12,
|
| 813 |
+
13,
|
| 814 |
+
14,
|
| 815 |
+
15,
|
| 816 |
+
16,
|
| 817 |
+
17,
|
| 818 |
+
18,
|
| 819 |
+
19,
|
| 820 |
+
20,
|
| 821 |
+
21,
|
| 822 |
+
22,
|
| 823 |
+
23,
|
| 824 |
+
24,
|
| 825 |
+
25,
|
| 826 |
+
26,
|
| 827 |
+
27,
|
| 828 |
+
28,
|
| 829 |
+
29,
|
| 830 |
+
30,
|
| 831 |
+
31,
|
| 832 |
+
32,
|
| 833 |
+
33,
|
| 834 |
+
34,
|
| 835 |
+
35,
|
| 836 |
+
36,
|
| 837 |
+
37,
|
| 838 |
+
38,
|
| 839 |
+
39
|
| 840 |
+
],
|
| 841 |
+
"modality_keys": [
|
| 842 |
+
"left_wrist_eef",
|
| 843 |
+
"right_wrist_eef",
|
| 844 |
+
"left_gripper_pos",
|
| 845 |
+
"right_gripper_pos",
|
| 846 |
+
"left_joint_pos",
|
| 847 |
+
"right_joint_pos"
|
| 848 |
+
],
|
| 849 |
+
"sin_cos_embedding_keys": null,
|
| 850 |
+
"mean_std_embedding_keys": null,
|
| 851 |
+
"action_configs": [
|
| 852 |
+
{
|
| 853 |
+
"rep": "RELATIVE",
|
| 854 |
+
"type": "EEF",
|
| 855 |
+
"format": "XYZ_ROT6D",
|
| 856 |
+
"state_key": "left_wrist_eef"
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"rep": "RELATIVE",
|
| 860 |
+
"type": "EEF",
|
| 861 |
+
"format": "XYZ_ROT6D",
|
| 862 |
+
"state_key": "right_wrist_eef"
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"rep": "ABSOLUTE",
|
| 866 |
+
"type": "NON_EEF",
|
| 867 |
+
"format": "DEFAULT",
|
| 868 |
+
"state_key": "left_gripper_pos"
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"rep": "ABSOLUTE",
|
| 872 |
+
"type": "NON_EEF",
|
| 873 |
+
"format": "DEFAULT",
|
| 874 |
+
"state_key": "right_gripper_pos"
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"rep": "RELATIVE",
|
| 878 |
+
"type": "NON_EEF",
|
| 879 |
+
"format": "DEFAULT",
|
| 880 |
+
"state_key": "left_joint_pos"
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"rep": "RELATIVE",
|
| 884 |
+
"type": "NON_EEF",
|
| 885 |
+
"format": "DEFAULT",
|
| 886 |
+
"state_key": "right_joint_pos"
|
| 887 |
+
}
|
| 888 |
+
]
|
| 889 |
+
},
|
| 890 |
+
"language": {
|
| 891 |
+
"delta_indices": [
|
| 892 |
+
0
|
| 893 |
+
],
|
| 894 |
+
"modality_keys": [
|
| 895 |
+
"annotation.sub_task"
|
| 896 |
+
],
|
| 897 |
+
"sin_cos_embedding_keys": null,
|
| 898 |
+
"mean_std_embedding_keys": null,
|
| 899 |
+
"action_configs": null
|
| 900 |
+
}
|
| 901 |
+
},
|
| 902 |
+
"oxe_droid_relative_eef_relative_joint": {
|
| 903 |
+
"video": {
|
| 904 |
+
"delta_indices": [
|
| 905 |
+
-15,
|
| 906 |
+
0
|
| 907 |
+
],
|
| 908 |
+
"modality_keys": [
|
| 909 |
+
"exterior_image_1_left",
|
| 910 |
+
"wrist_image_left"
|
| 911 |
+
],
|
| 912 |
+
"sin_cos_embedding_keys": null,
|
| 913 |
+
"mean_std_embedding_keys": null,
|
| 914 |
+
"action_configs": null
|
| 915 |
+
},
|
| 916 |
+
"state": {
|
| 917 |
+
"delta_indices": [
|
| 918 |
+
0
|
| 919 |
+
],
|
| 920 |
+
"modality_keys": [
|
| 921 |
+
"eef_9d",
|
| 922 |
+
"gripper_position",
|
| 923 |
+
"joint_position"
|
| 924 |
+
],
|
| 925 |
+
"sin_cos_embedding_keys": null,
|
| 926 |
+
"mean_std_embedding_keys": null,
|
| 927 |
+
"action_configs": null
|
| 928 |
+
},
|
| 929 |
+
"action": {
|
| 930 |
+
"delta_indices": [
|
| 931 |
+
0,
|
| 932 |
+
1,
|
| 933 |
+
2,
|
| 934 |
+
3,
|
| 935 |
+
4,
|
| 936 |
+
5,
|
| 937 |
+
6,
|
| 938 |
+
7,
|
| 939 |
+
8,
|
| 940 |
+
9,
|
| 941 |
+
10,
|
| 942 |
+
11,
|
| 943 |
+
12,
|
| 944 |
+
13,
|
| 945 |
+
14,
|
| 946 |
+
15,
|
| 947 |
+
16,
|
| 948 |
+
17,
|
| 949 |
+
18,
|
| 950 |
+
19,
|
| 951 |
+
20,
|
| 952 |
+
21,
|
| 953 |
+
22,
|
| 954 |
+
23,
|
| 955 |
+
24,
|
| 956 |
+
25,
|
| 957 |
+
26,
|
| 958 |
+
27,
|
| 959 |
+
28,
|
| 960 |
+
29,
|
| 961 |
+
30,
|
| 962 |
+
31,
|
| 963 |
+
32,
|
| 964 |
+
33,
|
| 965 |
+
34,
|
| 966 |
+
35,
|
| 967 |
+
36,
|
| 968 |
+
37,
|
| 969 |
+
38,
|
| 970 |
+
39
|
| 971 |
+
],
|
| 972 |
+
"modality_keys": [
|
| 973 |
+
"eef_9d",
|
| 974 |
+
"gripper_position",
|
| 975 |
+
"joint_position"
|
| 976 |
+
],
|
| 977 |
+
"sin_cos_embedding_keys": null,
|
| 978 |
+
"mean_std_embedding_keys": null,
|
| 979 |
+
"action_configs": [
|
| 980 |
+
{
|
| 981 |
+
"rep": "RELATIVE",
|
| 982 |
+
"type": "EEF",
|
| 983 |
+
"format": "XYZ_ROT6D",
|
| 984 |
+
"state_key": "eef_9d"
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"rep": "ABSOLUTE",
|
| 988 |
+
"type": "NON_EEF",
|
| 989 |
+
"format": "DEFAULT",
|
| 990 |
+
"state_key": "gripper_position"
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"rep": "RELATIVE",
|
| 994 |
+
"type": "NON_EEF",
|
| 995 |
+
"format": "DEFAULT",
|
| 996 |
+
"state_key": "joint_position"
|
| 997 |
+
}
|
| 998 |
+
]
|
| 999 |
+
},
|
| 1000 |
+
"language": {
|
| 1001 |
+
"delta_indices": [
|
| 1002 |
+
0
|
| 1003 |
+
],
|
| 1004 |
+
"modality_keys": [
|
| 1005 |
+
"annotation.language.language_instruction"
|
| 1006 |
+
],
|
| 1007 |
+
"sin_cos_embedding_keys": null,
|
| 1008 |
+
"mean_std_embedding_keys": null,
|
| 1009 |
+
"action_configs": null
|
| 1010 |
+
}
|
| 1011 |
+
},
|
| 1012 |
+
"new_embodiment": {
|
| 1013 |
+
"video": {
|
| 1014 |
+
"delta_indices": [
|
| 1015 |
+
0
|
| 1016 |
+
],
|
| 1017 |
+
"modality_keys": [
|
| 1018 |
+
"cam_left_head",
|
| 1019 |
+
"cam_left_wrist",
|
| 1020 |
+
"cam_right_wrist"
|
| 1021 |
+
],
|
| 1022 |
+
"sin_cos_embedding_keys": null,
|
| 1023 |
+
"mean_std_embedding_keys": null,
|
| 1024 |
+
"action_configs": null
|
| 1025 |
+
},
|
| 1026 |
+
"state": {
|
| 1027 |
+
"delta_indices": [
|
| 1028 |
+
0
|
| 1029 |
+
],
|
| 1030 |
+
"modality_keys": [
|
| 1031 |
+
"arm_left",
|
| 1032 |
+
"arm_right",
|
| 1033 |
+
"odometry"
|
| 1034 |
+
],
|
| 1035 |
+
"sin_cos_embedding_keys": null,
|
| 1036 |
+
"mean_std_embedding_keys": null,
|
| 1037 |
+
"action_configs": null
|
| 1038 |
+
},
|
| 1039 |
+
"action": {
|
| 1040 |
+
"delta_indices": [
|
| 1041 |
+
0,
|
| 1042 |
+
1,
|
| 1043 |
+
2,
|
| 1044 |
+
3,
|
| 1045 |
+
4,
|
| 1046 |
+
5,
|
| 1047 |
+
6,
|
| 1048 |
+
7,
|
| 1049 |
+
8,
|
| 1050 |
+
9,
|
| 1051 |
+
10,
|
| 1052 |
+
11,
|
| 1053 |
+
12,
|
| 1054 |
+
13,
|
| 1055 |
+
14,
|
| 1056 |
+
15
|
| 1057 |
+
],
|
| 1058 |
+
"modality_keys": [
|
| 1059 |
+
"arm_left",
|
| 1060 |
+
"arm_right",
|
| 1061 |
+
"odometry"
|
| 1062 |
+
],
|
| 1063 |
+
"sin_cos_embedding_keys": null,
|
| 1064 |
+
"mean_std_embedding_keys": null,
|
| 1065 |
+
"action_configs": [
|
| 1066 |
+
{
|
| 1067 |
+
"rep": "ABSOLUTE",
|
| 1068 |
+
"type": "NON_EEF",
|
| 1069 |
+
"format": "DEFAULT",
|
| 1070 |
+
"state_key": null
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"rep": "ABSOLUTE",
|
| 1074 |
+
"type": "NON_EEF",
|
| 1075 |
+
"format": "DEFAULT",
|
| 1076 |
+
"state_key": null
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"rep": "ABSOLUTE",
|
| 1080 |
+
"type": "NON_EEF",
|
| 1081 |
+
"format": "DEFAULT",
|
| 1082 |
+
"state_key": null
|
| 1083 |
+
}
|
| 1084 |
+
]
|
| 1085 |
+
},
|
| 1086 |
+
"language": {
|
| 1087 |
+
"delta_indices": [
|
| 1088 |
+
0
|
| 1089 |
+
],
|
| 1090 |
+
"modality_keys": [
|
| 1091 |
+
"annotation.human.task_description"
|
| 1092 |
+
],
|
| 1093 |
+
"sin_cos_embedding_keys": null,
|
| 1094 |
+
"mean_std_embedding_keys": null,
|
| 1095 |
+
"action_configs": null
|
| 1096 |
+
}
|
| 1097 |
+
}
|
| 1098 |
+
},
|
| 1099 |
+
"image_crop_size": [
|
| 1100 |
+
230,
|
| 1101 |
+
230
|
| 1102 |
+
],
|
| 1103 |
+
"image_target_size": [
|
| 1104 |
+
256,
|
| 1105 |
+
256
|
| 1106 |
+
],
|
| 1107 |
+
"use_albumentations": true,
|
| 1108 |
+
"random_rotation_angle": 0,
|
| 1109 |
+
"color_jitter_params": {
|
| 1110 |
+
"brightness": 0.3,
|
| 1111 |
+
"contrast": 0.4,
|
| 1112 |
+
"saturation": 0.5,
|
| 1113 |
+
"hue": 0.08
|
| 1114 |
+
},
|
| 1115 |
+
"shortest_image_edge": 256,
|
| 1116 |
+
"crop_fraction": 0.95,
|
| 1117 |
+
"letter_box_transform": false,
|
| 1118 |
+
"model_name": "nvidia/Cosmos-Reason2-2B",
|
| 1119 |
+
"model_type": "qwen",
|
| 1120 |
+
"formalize_language": true,
|
| 1121 |
+
"max_state_dim": 132,
|
| 1122 |
+
"max_action_dim": 132,
|
| 1123 |
+
"max_action_horizon": 40,
|
| 1124 |
+
"use_percentiles": false,
|
| 1125 |
+
"use_mean_std": false,
|
| 1126 |
+
"clip_outliers": true,
|
| 1127 |
+
"apply_sincos_state_encoding": false,
|
| 1128 |
+
"use_relative_action": true,
|
| 1129 |
+
"exclude_state": false,
|
| 1130 |
+
"state_dropout_prob": 0.2
|
| 1131 |
+
}
|
| 1132 |
+
}
|
*.pth/statistics.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
*.pth/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
*.pth/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fb246b0cdd309f00fa6859feca6c4a2ff98aa9e477a9a92cab2526b1e53b864e
|
| 3 |
+
size 6161
|
*.pth/wandb_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"project": "finetune-gr00t-n1d7", "run_id": "task0025_subtask_lang_final_layer_vlm_threecam_minmax_80k_gpu1_20260618_045240"}
|