Hugging Face's logo

RobotisSW
/
Task_802_groot_tuned_80k

Safetensors
Gr00tN1d7
Model card Files
xet
 Community
RobotisSW commited on
Commit
59973ba
·
verified ·
1 Parent(s): b7dd1c2

Upload folder using huggingface_hub

Browse files
Files changed (19) hide show
  1. config.json +97 -0
  2. embodiment_id.json +57 -0
  3. experiment_cfg/conf.yaml +233 -0
  4. experiment_cfg/config.yaml +231 -0
  5. experiment_cfg/dataset_statistics.json +1337 -0
  6. experiment_cfg/final_model_config.json +65 -0
  7. experiment_cfg/final_processor_config.json +0 -0
  8. model-00001-of-00003.safetensors +3 -0
  9. model-00002-of-00003.safetensors +3 -0
  10. model-00003-of-00003.safetensors +3 -0
  11. model.safetensors.index.json +0 -0
  12. optimizer.pt +3 -0
  13. processor_config.json +1129 -0
  14. rng_state.pth +3 -0
  15. scheduler.pt +3 -0
  16. statistics.json +0 -0
  17. trainer_state.json +0 -0
  18. training_args.bin +3 -0
  19. wandb_config.json +1 -0
config.json ADDED
@@ -0,0 +1,97 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "attn_supervision_layers": null,
11
+ "attn_supervision_loss": "kl",
12
+ "attn_supervision_margin": 0.9,
13
+ "attn_supervision_weight": 0.5,
14
+ "backbone_embedding_dim": 2048,
15
+ "backbone_trainable_params_fp32": true,
16
+ "color_jitter_params": {
17
+ "brightness": 0.3,
18
+ "contrast": 0.4,
19
+ "hue": 0.08,
20
+ "saturation": 0.5
21
+ },
22
+ "crop_fraction": 0.95,
23
+ "diffusion_model_cfg": {
24
+ "attention_head_dim": 48,
25
+ "dropout": 0.2,
26
+ "final_dropout": true,
27
+ "interleave_self_attention": true,
28
+ "norm_type": "ada_norm",
29
+ "num_attention_heads": 32,
30
+ "num_layers": 32,
31
+ "output_dim": 1024,
32
+ "positional_embeddings": null
33
+ },
34
+ "dtype": "float32",
35
+ "exclude_state": false,
36
+ "formalize_language": true,
37
+ "hidden_size": 1024,
38
+ "image_crop_size": [
39
+ 230,
40
+ 230
41
+ ],
42
+ "image_target_size": [
43
+ 256,
44
+ 256
45
+ ],
46
+ "letter_box_transform": false,
47
+ "load_bf16": false,
48
+ "max_action_dim": 132,
49
+ "max_num_embodiments": 32,
50
+ "max_seq_len": 1024,
51
+ "max_state_dim": 144,
52
+ "model_dtype": "bfloat16",
53
+ "model_name": "nvidia/Cosmos-Reason2-2B",
54
+ "model_type": "Gr00tN1d7",
55
+ "noise_beta_alpha": 1.5,
56
+ "noise_beta_beta": 1.0,
57
+ "noise_s": 0.999,
58
+ "num_inference_timesteps": 4,
59
+ "num_timestep_buckets": 1000,
60
+ "random_history_crop": true,
61
+ "random_rotation_angle": 0,
62
+ "reproject_vision": false,
63
+ "rtc_ramp_rate": 6.0,
64
+ "select_layer": 16,
65
+ "shortest_image_edge": 256,
66
+ "state_dropout_prob": 0.2,
67
+ "state_gaussian_noise_std": 0.0,
68
+ "transformers_version": "4.57.3",
69
+ "ttrtc_enabled": false,
70
+ "ttrtc_max_delay_steps": 10,
71
+ "ttrtc_policy_hz": 15.0,
72
+ "ttrtc_zero_delay_probability": 0.1,
73
+ "tune_diffusion_model": true,
74
+ "tune_linear": true,
75
+ "tune_llm": false,
76
+ "tune_projector": true,
77
+ "tune_top_llm_layers": 0,
78
+ "tune_visual": false,
79
+ "tune_vlln": true,
80
+ "use_albumentations": true,
81
+ "use_alternate_vl_dit": true,
82
+ "use_flash_attention": true,
83
+ "use_future_tokens": false,
84
+ "use_mean_std": false,
85
+ "use_percentiles": true,
86
+ "use_vl_self_attention": true,
87
+ "use_vlln": true,
88
+ "vision_kl_only": false,
89
+ "vl_self_attention_cfg": {
90
+ "attention_head_dim": 64,
91
+ "dropout": 0.2,
92
+ "final_dropout": true,
93
+ "num_attention_heads": 32,
94
+ "num_layers": 4,
95
+ "positional_embeddings": null
96
+ }
97
+ }
embodiment_id.json ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "robocasa_gr1_tabletop": 10,
55
+ "new_embodiment": 10,
56
+ "unitree_g1_sonic": 11
57
+ }
experiment_cfg/conf.yaml ADDED
@@ -0,0 +1,233 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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: null
18
+ image_target_size: null
19
+ shortest_image_edge: 256
20
+ crop_fraction: 0.9
21
+ random_rotation_angle: null
22
+ color_jitter_params: null
23
+ use_albumentations_transforms: true
24
+ letter_box_transform: true
25
+ extra_augmentation_config: null
26
+ attn_supervision_layers: null
27
+ vision_kl_only: false
28
+ attn_supervision_weight: 0.5
29
+ attn_supervision_loss: kl
30
+ attn_supervision_margin: 0.9
31
+ formalize_language: true
32
+ apply_sincos_state_encoding: false
33
+ use_percentiles: true
34
+ use_relative_action: true
35
+ max_state_dim: 132
36
+ max_action_dim: 132
37
+ action_horizon: 40
38
+ hidden_size: 1024
39
+ input_embedding_dim: 1536
40
+ state_history_length: 1
41
+ add_pos_embed: true
42
+ attn_dropout: 0.2
43
+ use_vlln: true
44
+ max_seq_len: 1024
45
+ use_alternate_vl_dit: true
46
+ attend_text_every_n_blocks: 2
47
+ diffusion_model_cfg:
48
+ positional_embeddings: null
49
+ num_layers: 16
50
+ num_attention_heads: 32
51
+ attention_head_dim: 48
52
+ norm_type: ada_norm
53
+ dropout: 0.2
54
+ final_dropout: true
55
+ output_dim: 1024
56
+ interleave_self_attention: true
57
+ num_inference_timesteps: 4
58
+ noise_beta_alpha: 1.5
59
+ noise_beta_beta: 1.0
60
+ noise_s: 0.999
61
+ num_timestep_buckets: 1000
62
+ ttrtc_enabled: false
63
+ ttrtc_max_delay_steps: 10
64
+ ttrtc_zero_delay_probability: 0.1
65
+ ttrtc_policy_hz: 15.0
66
+ tune_projector: true
67
+ tune_diffusion_model: true
68
+ tune_vlln: true
69
+ state_dropout_prob: 0.2
70
+ exclude_state: false
71
+ use_mean_std: false
72
+ max_num_embodiments: 32
73
+ data:
74
+ datasets:
75
+ - dataset_paths:
76
+ - /home/ubuntu/mission_reasoning/tuning/data/802_mcap_lerobot_v21
77
+ embodiment_tag: new_embodiment
78
+ mix_ratio: 1.0
79
+ dataset_type: physical_embodiment
80
+ val_dataset_path: null
81
+ modality_configs:
82
+ new_embodiment:
83
+ video:
84
+ delta_indices:
85
+ - 0
86
+ modality_keys:
87
+ - cam_left_head
88
+ sin_cos_embedding_keys: null
89
+ mean_std_embedding_keys: null
90
+ action_configs: null
91
+ state:
92
+ delta_indices:
93
+ - 0
94
+ modality_keys:
95
+ - left_arm
96
+ - right_arm
97
+ - left_hand
98
+ - right_hand
99
+ - tactile_left
100
+ - tactile_right
101
+ sin_cos_embedding_keys: null
102
+ mean_std_embedding_keys: null
103
+ action_configs: null
104
+ action:
105
+ delta_indices:
106
+ - 0
107
+ - 1
108
+ - 2
109
+ - 3
110
+ - 4
111
+ - 5
112
+ - 6
113
+ - 7
114
+ - 8
115
+ - 9
116
+ - 10
117
+ - 11
118
+ - 12
119
+ - 13
120
+ - 14
121
+ - 15
122
+ modality_keys:
123
+ - left_arm
124
+ - right_arm
125
+ - left_hand
126
+ - right_hand
127
+ sin_cos_embedding_keys: null
128
+ mean_std_embedding_keys: null
129
+ action_configs:
130
+ - rep: ABSOLUTE
131
+ type: NON_EEF
132
+ format: DEFAULT
133
+ state_key: null
134
+ - rep: ABSOLUTE
135
+ type: NON_EEF
136
+ format: DEFAULT
137
+ state_key: null
138
+ - rep: ABSOLUTE
139
+ type: NON_EEF
140
+ format: DEFAULT
141
+ state_key: null
142
+ - rep: ABSOLUTE
143
+ type: NON_EEF
144
+ format: DEFAULT
145
+ state_key: null
146
+ language:
147
+ delta_indices:
148
+ - 0
149
+ modality_keys:
150
+ - task
151
+ sin_cos_embedding_keys: null
152
+ mean_std_embedding_keys: null
153
+ action_configs: null
154
+ download_cache: false
155
+ shard_size: 256
156
+ episode_sampling_rate: 0.1
157
+ num_shards_per_epoch: 100000
158
+ ds_weights_alpha: null
159
+ override_pretraining_statistics: true
160
+ mode: single_turn
161
+ random_chop: 0.0
162
+ mock_dataset_mode: false
163
+ shuffle: true
164
+ seed: 42
165
+ multiprocessing_context: fork
166
+ allow_padding: false
167
+ subsample_ratio: 1.0
168
+ image_crop_size:
169
+ - 244
170
+ - 244
171
+ image_target_size:
172
+ - 224
173
+ - 224
174
+ training:
175
+ output_dir: /home/ubuntu/mission_reasoning/tuning/checkpoints/action
176
+ experiment_name: 802_v21_frozen_vlm_action_100k_20261104
177
+ max_steps: 100000
178
+ global_batch_size: 128
179
+ per_gpu_batch_size: null
180
+ gradient_accumulation_steps: 1
181
+ learning_rate: 0.0001
182
+ lr_scheduler_type: cosine
183
+ weight_decay: 1.0e-05
184
+ warmup_ratio: 0.05
185
+ warmup_steps: 0
186
+ max_grad_norm: 1.0
187
+ optim: adamw_torch
188
+ start_from_checkpoint: /home/ubuntu/mission_reasoning/tuning/checkpoints/base-gr00t-state144
189
+ skip_weight_loading: false
190
+ tf32: true
191
+ fp16: false
192
+ bf16: true
193
+ eval_bf16: true
194
+ logging_steps: 10
195
+ save_steps: 10000
196
+ save_total_limit: 2
197
+ save_vl_model: false
198
+ save_only_model: false
199
+ resume_from_checkpoint: false
200
+ upload_checkpoints: false
201
+ upload_every: 1000
202
+ upload_last_n_checkpoints: 5
203
+ max_concurrent_uploads: 2
204
+ eval_strategy: 'no'
205
+ eval_steps: 500
206
+ eval_set_split_ratio: 0.1
207
+ eval_batch_size: 2
208
+ save_best_eval_metric_name: ''
209
+ save_best_eval_metric_greater_is_better: true
210
+ deepspeed_stage: 2
211
+ gradient_checkpointing: false
212
+ transformers_trust_remote_code: true
213
+ transformers_local_files_only: false
214
+ transformers_cache_dir: null
215
+ transformers_access_token: null
216
+ use_ddp: false
217
+ ddp_bucket_cap_mb: 100
218
+ num_gpus: 1
219
+ dataloader_num_workers: 2
220
+ remove_unused_columns: false
221
+ use_wandb: false
222
+ wandb_project: finetune-gr00t-n1d7
223
+ enable_profiling: false
224
+ max_retries: 3
225
+ assert_loss_less_than: null
226
+ add_rl_callback: false
227
+ enable_open_loop_eval: false
228
+ open_loop_eval_traj_ids:
229
+ - 0
230
+ open_loop_eval_steps_per_traj: 100
231
+ open_loop_eval_plot_indices: null
232
+ max_steps: 100000
233
+ save_steps: 10000
experiment_cfg/config.yaml ADDED
@@ -0,0 +1,231 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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: null
18
+ image_target_size: null
19
+ shortest_image_edge: 256
20
+ crop_fraction: 0.9
21
+ random_rotation_angle: null
22
+ color_jitter_params: null
23
+ use_albumentations_transforms: true
24
+ letter_box_transform: true
25
+ extra_augmentation_config: null
26
+ attn_supervision_layers: null
27
+ vision_kl_only: false
28
+ attn_supervision_weight: 0.5
29
+ attn_supervision_loss: kl
30
+ attn_supervision_margin: 0.9
31
+ formalize_language: true
32
+ apply_sincos_state_encoding: false
33
+ use_percentiles: true
34
+ use_relative_action: true
35
+ max_state_dim: 132
36
+ max_action_dim: 132
37
+ action_horizon: 40
38
+ hidden_size: 1024
39
+ input_embedding_dim: 1536
40
+ state_history_length: 1
41
+ add_pos_embed: true
42
+ attn_dropout: 0.2
43
+ use_vlln: true
44
+ max_seq_len: 1024
45
+ use_alternate_vl_dit: true
46
+ attend_text_every_n_blocks: 2
47
+ diffusion_model_cfg:
48
+ positional_embeddings: null
49
+ num_layers: 16
50
+ num_attention_heads: 32
51
+ attention_head_dim: 48
52
+ norm_type: ada_norm
53
+ dropout: 0.2
54
+ final_dropout: true
55
+ output_dim: 1024
56
+ interleave_self_attention: true
57
+ num_inference_timesteps: 4
58
+ noise_beta_alpha: 1.5
59
+ noise_beta_beta: 1.0
60
+ noise_s: 0.999
61
+ num_timestep_buckets: 1000
62
+ ttrtc_enabled: false
63
+ ttrtc_max_delay_steps: 10
64
+ ttrtc_zero_delay_probability: 0.1
65
+ ttrtc_policy_hz: 15.0
66
+ tune_projector: true
67
+ tune_diffusion_model: true
68
+ tune_vlln: true
69
+ state_dropout_prob: 0.2
70
+ exclude_state: false
71
+ use_mean_std: false
72
+ max_num_embodiments: 32
73
+ data:
74
+ datasets:
75
+ - dataset_paths:
76
+ - /home/ubuntu/mission_reasoning/tuning/data/802_mcap_lerobot_v21
77
+ embodiment_tag: new_embodiment
78
+ mix_ratio: 1.0
79
+ dataset_type: physical_embodiment
80
+ val_dataset_path: null
81
+ modality_configs:
82
+ new_embodiment:
83
+ video:
84
+ delta_indices:
85
+ - 0
86
+ modality_keys:
87
+ - cam_left_head
88
+ sin_cos_embedding_keys: null
89
+ mean_std_embedding_keys: null
90
+ action_configs: null
91
+ state:
92
+ delta_indices:
93
+ - 0
94
+ modality_keys:
95
+ - left_arm
96
+ - right_arm
97
+ - left_hand
98
+ - right_hand
99
+ - tactile_left
100
+ - tactile_right
101
+ sin_cos_embedding_keys: null
102
+ mean_std_embedding_keys: null
103
+ action_configs: null
104
+ action:
105
+ delta_indices:
106
+ - 0
107
+ - 1
108
+ - 2
109
+ - 3
110
+ - 4
111
+ - 5
112
+ - 6
113
+ - 7
114
+ - 8
115
+ - 9
116
+ - 10
117
+ - 11
118
+ - 12
119
+ - 13
120
+ - 14
121
+ - 15
122
+ modality_keys:
123
+ - left_arm
124
+ - right_arm
125
+ - left_hand
126
+ - right_hand
127
+ sin_cos_embedding_keys: null
128
+ mean_std_embedding_keys: null
129
+ action_configs:
130
+ - rep: absolute
131
+ type: non_eef
132
+ format: default
133
+ state_key: null
134
+ - rep: absolute
135
+ type: non_eef
136
+ format: default
137
+ state_key: null
138
+ - rep: absolute
139
+ type: non_eef
140
+ format: default
141
+ state_key: null
142
+ - rep: absolute
143
+ type: non_eef
144
+ format: default
145
+ state_key: null
146
+ language:
147
+ delta_indices:
148
+ - 0
149
+ modality_keys:
150
+ - task
151
+ sin_cos_embedding_keys: null
152
+ mean_std_embedding_keys: null
153
+ action_configs: null
154
+ download_cache: false
155
+ shard_size: 256
156
+ episode_sampling_rate: 0.1
157
+ num_shards_per_epoch: 100000
158
+ ds_weights_alpha: null
159
+ override_pretraining_statistics: true
160
+ mode: single_turn
161
+ random_chop: 0.0
162
+ mock_dataset_mode: false
163
+ shuffle: true
164
+ seed: 42
165
+ multiprocessing_context: fork
166
+ allow_padding: false
167
+ subsample_ratio: 1.0
168
+ image_crop_size:
169
+ - 244
170
+ - 244
171
+ image_target_size:
172
+ - 224
173
+ - 224
174
+ training:
175
+ output_dir: /home/ubuntu/mission_reasoning/tuning/checkpoints/action
176
+ experiment_name: 802_v21_frozen_vlm_action_100k_20261104
177
+ max_steps: 100000
178
+ global_batch_size: 128
179
+ per_gpu_batch_size: null
180
+ gradient_accumulation_steps: 1
181
+ learning_rate: 0.0001
182
+ lr_scheduler_type: cosine
183
+ weight_decay: 1.0e-05
184
+ warmup_ratio: 0.05
185
+ warmup_steps: 0
186
+ max_grad_norm: 1.0
187
+ optim: adamw_torch
188
+ start_from_checkpoint: /home/ubuntu/mission_reasoning/tuning/checkpoints/base-gr00t-state144
189
+ skip_weight_loading: false
190
+ tf32: true
191
+ fp16: false
192
+ bf16: true
193
+ eval_bf16: true
194
+ logging_steps: 10
195
+ save_steps: 10000
196
+ save_total_limit: 2
197
+ save_vl_model: false
198
+ save_only_model: false
199
+ resume_from_checkpoint: false
200
+ upload_checkpoints: false
201
+ upload_every: 1000
202
+ upload_last_n_checkpoints: 5
203
+ max_concurrent_uploads: 2
204
+ eval_strategy: 'no'
205
+ eval_steps: 500
206
+ eval_set_split_ratio: 0.1
207
+ eval_batch_size: 2
208
+ save_best_eval_metric_name: ''
209
+ save_best_eval_metric_greater_is_better: true
210
+ deepspeed_stage: 2
211
+ gradient_checkpointing: false
212
+ transformers_trust_remote_code: true
213
+ transformers_local_files_only: false
214
+ transformers_cache_dir: null
215
+ transformers_access_token: null
216
+ use_ddp: false
217
+ ddp_bucket_cap_mb: 100
218
+ num_gpus: 1
219
+ dataloader_num_workers: 2
220
+ remove_unused_columns: false
221
+ use_wandb: false
222
+ wandb_project: finetune-gr00t-n1d7
223
+ enable_profiling: false
224
+ max_retries: 3
225
+ assert_loss_less_than: null
226
+ add_rl_callback: false
227
+ enable_open_loop_eval: false
228
+ open_loop_eval_traj_ids:
229
+ - 0
230
+ open_loop_eval_steps_per_traj: 100
231
+ open_loop_eval_plot_indices: null
experiment_cfg/dataset_statistics.json ADDED
@@ -0,0 +1,1337 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "new_embodiment": {
3
+ "state": {
4
+ "left_arm": {
5
+ "min": [
6
+ -0.13344433903694153,
7
+ -1.1984225238848012e-05,
8
+ -0.1426841765642166,
9
+ -2.6476387977600098,
10
+ -0.2070874124765396,
11
+ 0.8743570446968079,
12
+ -0.1196235716342926
13
+ ],
14
+ "max": [
15
+ 0.08895890414714813,
16
+ 1.1984225238848012e-05,
17
+ -0.052143361419439316,
18
+ -2.574007749557495,
19
+ -0.04140549525618553,
20
+ 1.0929492712020874,
21
+ -0.009163624607026577
22
+ ],
23
+ "mean": [
24
+ -0.054605789482593536,
25
+ 3.2164207408591494e-10,
26
+ -0.10622034966945648,
27
+ -2.6099092960357666,
28
+ -0.19246385991573334,
29
+ 1.0605239868164062,
30
+ -0.11079040914773941
31
+ ],
32
+ "std": [
33
+ 0.05639733374118805,
34
+ 1.8625652842274576e-07,
35
+ 0.013397497124969926,
36
+ 0.016928216442453382,
37
+ 0.032859351485967636,
38
+ 0.07246726751327515,
39
+ 0.028111889958381653
40
+ ],
41
+ "q01": [
42
+ -0.13344433903694153,
43
+ 0.0,
44
+ -0.1426602154970169,
45
+ -2.6476387977600098,
46
+ -0.2070874124765396,
47
+ 0.8743570446968079,
48
+ -0.1196235716342926
49
+ ],
50
+ "q99": [
51
+ 0.08895890414714813,
52
+ 0.0,
53
+ -0.05215534567832947,
54
+ -2.574007749557495,
55
+ -0.041417479515075684,
56
+ 1.0929492712020874,
57
+ -0.009163624607026577
58
+ ]
59
+ },
60
+ "right_arm": {
61
+ "min": [
62
+ -0.5563796162605286,
63
+ -0.6961036920547485,
64
+ -0.5945613384246826,
65
+ -2.9193332195281982,
66
+ -0.6392265558242798,
67
+ -1.1935688257217407,
68
+ -0.5612978935241699
69
+ ],
70
+ "max": [
71
+ 0.37515416741371155,
72
+ 0.00017976337403524667,
73
+ 0.5537310838699341,
74
+ -1.2355976104736328,
75
+ 1.94078528881073,
76
+ 1.5700054168701172,
77
+ 1.39189875125885
78
+ ],
79
+ "mean": [
80
+ -0.12533333897590637,
81
+ -0.07734235376119614,
82
+ 0.025785373523831367,
83
+ -2.2696666717529297,
84
+ 0.6330434083938599,
85
+ 0.6639376878738403,
86
+ 0.40092214941978455
87
+ ],
88
+ "std": [
89
+ 0.1421017348766327,
90
+ 0.10486652702093124,
91
+ 0.15915478765964508,
92
+ 0.36544618010520935,
93
+ 0.5626290440559387,
94
+ 0.5100393891334534,
95
+ 0.3716941177845001
96
+ ],
97
+ "q01": [
98
+ -0.4804684311151505,
99
+ -0.5552385103702545,
100
+ -0.3342618536949158,
101
+ -2.7593319416046143,
102
+ -0.34039487838745114,
103
+ -0.9329884278774261,
104
+ -0.35206337392330167
105
+ ],
106
+ "q99": [
107
+ 0.19797076195478402,
108
+ 4.793690095539205e-05,
109
+ 0.38389068841934204,
110
+ -1.4228390455245972,
111
+ 1.6618309140205374,
112
+ 1.365040357112884,
113
+ 1.1319465708732601
114
+ ]
115
+ },
116
+ "left_hand": {
117
+ "min": [
118
+ -0.7731263041496277,
119
+ -3.1461946964263916,
120
+ -3.1247189044952393,
121
+ -0.251572847366333,
122
+ -0.1871456503868103,
123
+ 1.1382137537002563,
124
+ 0.5875146389007568,
125
+ 0.7439807057380676,
126
+ -0.10124272853136063,
127
+ 1.274738073348999,
128
+ 0.5737088322639465,
129
+ 0.6442719101905823,
130
+ 0.01073786523193121,
131
+ 1.1550875902175903,
132
+ 0.6059224009513855,
133
+ 0.7133010625839233,
134
+ -0.11351457983255386,
135
+ 0.977145791053772,
136
+ 0.5675728917121887,
137
+ 0.849825382232666
138
+ ],
139
+ "max": [
140
+ 0.09357282519340515,
141
+ 1.572330355644226,
142
+ 0.5154175162315369,
143
+ 38.262081146240234,
144
+ -0.15493206679821014,
145
+ 1.2701361179351807,
146
+ 0.6442719101905823,
147
+ 0.7593204975128174,
148
+ -0.07209710031747818,
149
+ 1.352971076965332,
150
+ 0.6135923266410828,
151
+ 0.6810874938964844,
152
+ 0.04448544234037399,
153
+ 1.2394565343856812,
154
+ 0.6458058953285217,
155
+ 0.7347767949104309,
156
+ 0.0951068103313446,
157
+ 1.0753204822540283,
158
+ 0.630466103553772,
159
+ 0.8774369955062866
160
+ ],
161
+ "mean": [
162
+ -0.6870552897453308,
163
+ 1.5080620050430298,
164
+ -0.7994865775108337,
165
+ -0.21608512103557587,
166
+ -0.16863705217838287,
167
+ 1.197111964225769,
168
+ 0.619471549987793,
169
+ 0.7555537819862366,
170
+ -0.08983048796653748,
171
+ 1.3010039329528809,
172
+ 0.5898144245147705,
173
+ 0.6669421195983887,
174
+ 0.0253777913749218,
175
+ 1.1902176141738892,
176
+ 0.6249401569366455,
177
+ 0.7263370156288147,
178
+ 0.02219611033797264,
179
+ 1.009318470954895,
180
+ 0.6186517477035522,
181
+ 0.8608671426773071
182
+ ],
183
+ "std": [
184
+ 0.04533351585268959,
185
+ 0.03951200470328595,
186
+ 0.6161787509918213,
187
+ 0.28086239099502563,
188
+ 0.009025214239954948,
189
+ 0.034269168972969055,
190
+ 0.011652555316686035,
191
+ 0.002153336536152089,
192
+ 0.007937277667224353,
193
+ 0.02600982226431523,
194
+ 0.012295232154430052,
195
+ 0.011742408387362034,
196
+ 0.008401039056479931,
197
+ 0.029938405379652457,
198
+ 0.01184874307364134,
199
+ 0.006560818757858531,
200
+ 0.06808660179376602,
201
+ 0.02841388061642818,
202
+ 0.0053639090620002106,
203
+ 0.008137298747900345
204
+ ],
205
+ "q01": [
206
+ -0.7731263041496277,
207
+ 1.492563247680664,
208
+ -1.162757396697998,
209
+ -0.25003886222839355,
210
+ -0.18561168015003204,
211
+ 1.1596894264221191,
212
+ 0.5997864603996277,
213
+ 0.7470486164093018,
214
+ -0.09970875084400177,
215
+ 1.2762720584869385,
216
+ 0.5737088322639465,
217
+ 0.6442719101905823,
218
+ 0.013805827125906944,
219
+ 1.1596894264221191,
220
+ 0.6105243563652039,
221
+ 0.7133010625839233,
222
+ -0.11198059469461441,
223
+ 0.983281672000885,
224
+ 0.6059224009513855,
225
+ 0.849825382232666
226
+ ],
227
+ "q99": [
228
+ -0.6243301630020142,
229
+ 1.5692623853683472,
230
+ 0.5138835906982422,
231
+ -0.20248545706272125,
232
+ -0.15493206679821014,
233
+ 1.2701361179351807,
234
+ 0.6442719101905823,
235
+ 0.7593204975128174,
236
+ -0.07209710031747818,
237
+ 1.3514370918273926,
238
+ 0.6105243563652039,
239
+ 0.676485538482666,
240
+ 0.04448544234037399,
241
+ 1.2379225492477417,
242
+ 0.6442719101905823,
243
+ 0.731708824634552,
244
+ 0.0951068103313446,
245
+ 1.06918466091156,
246
+ 0.6258641481399536,
247
+ 0.8774369955062866
248
+ ]
249
+ },
250
+ "right_hand": {
251
+ "min": [
252
+ -0.02454369328916073,
253
+ -1.7027186155319214,
254
+ -0.4724660813808441,
255
+ -0.01840776950120926,
256
+ 0.01073786523193121,
257
+ -0.02761165425181389,
258
+ -0.019941750913858414,
259
+ -0.01840776950120926,
260
+ -0.02761165425181389,
261
+ -0.006135923322290182,
262
+ -0.01073786523193121,
263
+ -0.015339807607233524,
264
+ -0.06749515235424042,
265
+ -0.019941750913858414,
266
+ -0.016873788088560104,
267
+ -0.026077672839164734,
268
+ -0.1564660370349884,
269
+ -0.03988350182771683,
270
+ -0.02147573046386242,
271
+ -0.029145635664463043
272
+ ],
273
+ "max": [
274
+ 0.777728259563446,
275
+ -1.414330244064331,
276
+ 1.116737961769104,
277
+ 0.23316508531570435,
278
+ 0.42031073570251465,
279
+ 40.84070587158203,
280
+ 41.233402252197266,
281
+ 41.626102447509766,
282
+ 0.1257864236831665,
283
+ 1.3345632553100586,
284
+ 0.5967185497283936,
285
+ 0.6841554045677185,
286
+ 0.03067961521446705,
287
+ 1.2302526235580444,
288
+ 0.6473398804664612,
289
+ 0.7593204975128174,
290
+ 0.01840776950120926,
291
+ 0.9970875382423401,
292
+ 0.6273981332778931,
293
+ 0.9019806981086731
294
+ ],
295
+ "mean": [
296
+ 0.38932913541793823,
297
+ -1.5264192819595337,
298
+ 0.485247939825058,
299
+ 0.08364326506853104,
300
+ 0.10003503412008286,
301
+ 0.6452912092208862,
302
+ 0.32987913489341736,
303
+ 0.4391363263130188,
304
+ 0.056282829493284225,
305
+ 0.7282194495201111,
306
+ 0.30848750472068787,
307
+ 0.38219860196113586,
308
+ -0.006870318204164505,
309
+ 0.6751349568367004,
310
+ 0.3424829840660095,
311
+ 0.42355895042419434,
312
+ -0.06138969585299492,
313
+ 0.5465292930603027,
314
+ 0.3390616178512573,
315
+ 0.5172824263572693
316
+ ],
317
+ "std": [
318
+ 0.22187680006027222,
319
+ 0.04463353380561108,
320
+ 0.28433048725128174,
321
+ 0.05792798474431038,
322
+ 0.03868968039751053,
323
+ 0.3879151940345764,
324
+ 0.2708965241909027,
325
+ 0.3154506981372833,
326
+ 0.023756863549351692,
327
+ 0.3665197193622589,
328
+ 0.1471555233001709,
329
+ 0.19984331727027893,
330
+ 0.012469855137169361,
331
+ 0.33706364035606384,
332
+ 0.16844700276851654,
333
+ 0.22270169854164124,
334
+ 0.03866128996014595,
335
+ 0.28351956605911255,
336
+ 0.17650021612644196,
337
+ 0.2785572111606598
338
+ ],
339
+ "q01": [
340
+ 0.08130098134279251,
341
+ -1.6643691062927246,
342
+ -0.16260196268558502,
343
+ -0.0015339808305725455,
344
+ 0.03681553900241852,
345
+ 0.16566991806030273,
346
+ 0.08436894416809082,
347
+ 0.10124272853136063,
348
+ 0.007669903803616762,
349
+ 0.17947575449943542,
350
+ 0.0782330185174942,
351
+ 0.08590292185544968,
352
+ -0.041417479515075684,
353
+ 0.17180584371089935,
354
+ 0.08436894416809082,
355
+ 0.09357282519340515,
356
+ -0.14572817087173462,
357
+ 0.12885437905788422,
358
+ 0.07976700365543365,
359
+ 0.11965049803256989
360
+ ],
361
+ "q99": [
362
+ 0.7593204975128174,
363
+ -1.4649516344070435,
364
+ 1.1075341701507568,
365
+ 0.22242721915245056,
366
+ 0.17947575449943542,
367
+ 1.1995729207992554,
368
+ 0.630466103553772,
369
+ 0.7746602892875671,
370
+ 0.11351457983255386,
371
+ 1.2946797609329224,
372
+ 0.5568350553512573,
373
+ 0.6580777764320374,
374
+ 0.013805827125906944,
375
+ 1.2041748762130737,
376
+ 0.6151263117790222,
377
+ 0.7271068692207336,
378
+ 0.0015339808305725455,
379
+ 0.9924855828285217,
380
+ 0.6179181647300604,
381
+ 0.8958447575569153
382
+ ]
383
+ },
384
+ "tactile_left": {
385
+ "min": [
386
+ 28.0,
387
+ 31.0,
388
+ 10.0,
389
+ 29.0,
390
+ 33.0,
391
+ 13.0,
392
+ 20.0,
393
+ 23.0,
394
+ 4.0,
395
+ 0.0,
396
+ 0.0,
397
+ 0.0,
398
+ 0.0,
399
+ 0.0,
400
+ 0.0,
401
+ 0.0,
402
+ 0.0,
403
+ 0.0,
404
+ 0.0,
405
+ 0.0,
406
+ 0.0,
407
+ 0.0,
408
+ 0.0,
409
+ 0.0,
410
+ 0.0,
411
+ 0.0,
412
+ 0.0,
413
+ 0.0,
414
+ 0.0,
415
+ 0.0,
416
+ 0.0,
417
+ 0.0,
418
+ 0.0,
419
+ 0.0,
420
+ 0.0,
421
+ 0.0,
422
+ 0.0,
423
+ 7.0,
424
+ 3.0,
425
+ 0.0,
426
+ 7.0,
427
+ 2.0,
428
+ 0.0,
429
+ 21.0,
430
+ 6.0
431
+ ],
432
+ "max": [
433
+ 161.0,
434
+ 167.0,
435
+ 158.0,
436
+ 250.0,
437
+ 233.0,
438
+ 162.0,
439
+ 164.0,
440
+ 175.0,
441
+ 165.0,
442
+ 154.0,
443
+ 162.0,
444
+ 157.0,
445
+ 150.0,
446
+ 155.0,
447
+ 151.0,
448
+ 152.0,
449
+ 156.0,
450
+ 156.0,
451
+ 156.0,
452
+ 168.0,
453
+ 164.0,
454
+ 156.0,
455
+ 160.0,
456
+ 158.0,
457
+ 159.0,
458
+ 168.0,
459
+ 158.0,
460
+ 155.0,
461
+ 163.0,
462
+ 158.0,
463
+ 152.0,
464
+ 155.0,
465
+ 153.0,
466
+ 157.0,
467
+ 163.0,
468
+ 157.0,
469
+ 132.0,
470
+ 137.0,
471
+ 135.0,
472
+ 129.0,
473
+ 132.0,
474
+ 130.0,
475
+ 134.0,
476
+ 94.0,
477
+ 6.0
478
+ ],
479
+ "mean": [
480
+ 92.38203430175781,
481
+ 98.35807037353516,
482
+ 80.30078125,
483
+ 92.51570892333984,
484
+ 99.26548767089844,
485
+ 83.32854461669922,
486
+ 87.38018798828125,
487
+ 95.16789245605469,
488
+ 77.56182861328125,
489
+ 52.724822998046875,
490
+ 64.72637939453125,
491
+ 52.99644470214844,
492
+ 52.614097595214844,
493
+ 62.246917724609375,
494
+ 50.91588592529297,
495
+ 46.025978088378906,
496
+ 49.5438346862793,
497
+ 49.032623291015625,
498
+ 47.69234848022461,
499
+ 58.32101821899414,
500
+ 50.65286636352539,
501
+ 47.56197738647461,
502
+ 52.01638412475586,
503
+ 42.6334342956543,
504
+ 41.131752014160156,
505
+ 44.54442596435547,
506
+ 34.65938949584961,
507
+ 65.46189880371094,
508
+ 80.37161254882812,
509
+ 71.0626449584961,
510
+ 65.15083312988281,
511
+ 77.52117919921875,
512
+ 69.10393524169922,
513
+ 59.35449981689453,
514
+ 70.15735626220703,
515
+ 60.06256103515625,
516
+ 58.929603576660156,
517
+ 69.57051086425781,
518
+ 63.67304992675781,
519
+ 58.07906723022461,
520
+ 67.63542175292969,
521
+ 62.26048278808594,
522
+ 53.952911376953125,
523
+ 72.04816436767578,
524
+ 6.0
525
+ ],
526
+ "std": [
527
+ 30.724609375,
528
+ 29.91054916381836,
529
+ 33.873172760009766,
530
+ 30.003475189208984,
531
+ 29.369510650634766,
532
+ 34.47971725463867,
533
+ 35.43425750732422,
534
+ 34.894832611083984,
535
+ 39.897178649902344,
536
+ 38.788604736328125,
537
+ 37.80418395996094,
538
+ 40.46141052246094,
539
+ 37.47084426879883,
540
+ 36.672908782958984,
541
+ 38.95347213745117,
542
+ 41.316158294677734,
543
+ 41.30397415161133,
544
+ 41.35035705566406,
545
+ 42.148712158203125,
546
+ 41.30094909667969,
547
+ 43.617000579833984,
548
+ 42.24308395385742,
549
+ 41.32605743408203,
550
+ 46.18037033081055,
551
+ 47.24040222167969,
552
+ 47.802101135253906,
553
+ 52.08993911743164,
554
+ 36.31819152832031,
555
+ 34.780765533447266,
556
+ 35.726505279541016,
557
+ 35.02424240112305,
558
+ 33.434967041015625,
559
+ 34.67414093017578,
560
+ 39.37534713745117,
561
+ 36.9697380065918,
562
+ 39.04301452636719,
563
+ 30.627878189086914,
564
+ 27.361858367919922,
565
+ 29.312007904052734,
566
+ 29.809783935546875,
567
+ 26.39284896850586,
568
+ 27.968704223632812,
569
+ 34.07364273071289,
570
+ 25.13787841796875,
571
+ 0.0
572
+ ],
573
+ "q01": [
574
+ 32.0,
575
+ 34.0,
576
+ 15.0,
577
+ 33.0,
578
+ 36.0,
579
+ 17.0,
580
+ 24.0,
581
+ 27.0,
582
+ 9.0,
583
+ 0.0,
584
+ 0.0,
585
+ 0.0,
586
+ 0.0,
587
+ 0.0,
588
+ 0.0,
589
+ 0.0,
590
+ 0.0,
591
+ 0.0,
592
+ 0.0,
593
+ 0.0,
594
+ 0.0,
595
+ 0.0,
596
+ 0.0,
597
+ 0.0,
598
+ 0.0,
599
+ 0.0,
600
+ 0.0,
601
+ 0.0,
602
+ 0.0,
603
+ 0.0,
604
+ 0.0,
605
+ 0.0,
606
+ 0.0,
607
+ 0.0,
608
+ 0.0,
609
+ 0.0,
610
+ 0.0,
611
+ 10.0,
612
+ 6.0,
613
+ 0.0,
614
+ 9.0,
615
+ 5.0,
616
+ 0.0,
617
+ 21.0,
618
+ 6.0
619
+ ],
620
+ "q99": [
621
+ 160.0,
622
+ 166.0,
623
+ 155.0,
624
+ 158.0,
625
+ 164.0,
626
+ 158.0,
627
+ 163.0,
628
+ 173.0,
629
+ 162.0,
630
+ 152.0,
631
+ 160.0,
632
+ 156.0,
633
+ 148.0,
634
+ 153.0,
635
+ 150.0,
636
+ 151.0,
637
+ 155.0,
638
+ 155.0,
639
+ 155.0,
640
+ 166.0,
641
+ 162.0,
642
+ 155.0,
643
+ 159.0,
644
+ 157.0,
645
+ 158.0,
646
+ 167.0,
647
+ 156.0,
648
+ 155.0,
649
+ 162.0,
650
+ 157.0,
651
+ 151.0,
652
+ 154.0,
653
+ 152.0,
654
+ 156.0,
655
+ 162.0,
656
+ 155.0,
657
+ 131.0,
658
+ 135.0,
659
+ 134.0,
660
+ 128.0,
661
+ 130.0,
662
+ 129.0,
663
+ 134.0,
664
+ 94.0,
665
+ 6.0
666
+ ]
667
+ },
668
+ "tactile_right": {
669
+ "min": [
670
+ 0.0,
671
+ 0.0,
672
+ 0.0,
673
+ 0.0,
674
+ 0.0,
675
+ 0.0,
676
+ 0.0,
677
+ 0.0,
678
+ 0.0,
679
+ 0.0,
680
+ 0.0,
681
+ 0.0,
682
+ 0.0,
683
+ 0.0,
684
+ 0.0,
685
+ 0.0,
686
+ 0.0,
687
+ 146.0,
688
+ 0.0,
689
+ 0.0,
690
+ 0.0,
691
+ 0.0,
692
+ 0.0,
693
+ 0.0,
694
+ 0.0,
695
+ 0.0,
696
+ 0.0,
697
+ 0.0,
698
+ 0.0,
699
+ 0.0,
700
+ 0.0,
701
+ 0.0,
702
+ 0.0,
703
+ 0.0,
704
+ 0.0,
705
+ 0.0,
706
+ 0.0,
707
+ 0.0,
708
+ 85.0,
709
+ 0.0,
710
+ 0.0,
711
+ 0.0,
712
+ 0.0,
713
+ 0.0,
714
+ 8.0
715
+ ],
716
+ "max": [
717
+ 255.0,
718
+ 255.0,
719
+ 255.0,
720
+ 255.0,
721
+ 255.0,
722
+ 255.0,
723
+ 252.0,
724
+ 255.0,
725
+ 255.0,
726
+ 255.0,
727
+ 255.0,
728
+ 255.0,
729
+ 255.0,
730
+ 255.0,
731
+ 255.0,
732
+ 255.0,
733
+ 255.0,
734
+ 255.0,
735
+ 255.0,
736
+ 255.0,
737
+ 255.0,
738
+ 255.0,
739
+ 255.0,
740
+ 232.0,
741
+ 255.0,
742
+ 255.0,
743
+ 255.0,
744
+ 235.0,
745
+ 255.0,
746
+ 255.0,
747
+ 174.0,
748
+ 240.0,
749
+ 255.0,
750
+ 179.0,
751
+ 230.0,
752
+ 226.0,
753
+ 185.0,
754
+ 255.0,
755
+ 121.0,
756
+ 184.0,
757
+ 205.0,
758
+ 246.0,
759
+ 179.0,
760
+ 255.0,
761
+ 9.0
762
+ ],
763
+ "mean": [
764
+ 92.95165252685547,
765
+ 105.3428955078125,
766
+ 82.301025390625,
767
+ 85.60537719726562,
768
+ 95.72203063964844,
769
+ 81.19490051269531,
770
+ 47.246124267578125,
771
+ 87.92813873291016,
772
+ 71.41304779052734,
773
+ 49.98362731933594,
774
+ 46.76289367675781,
775
+ 46.849674224853516,
776
+ 49.35315704345703,
777
+ 58.55553436279297,
778
+ 52.30145263671875,
779
+ 46.13134765625,
780
+ 71.51017761230469,
781
+ 238.06057739257812,
782
+ 33.010292053222656,
783
+ 36.43037414550781,
784
+ 27.28377914428711,
785
+ 26.809484481811523,
786
+ 34.30710220336914,
787
+ 23.306699752807617,
788
+ 24.876943588256836,
789
+ 27.127525329589844,
790
+ 26.53605079650879,
791
+ 30.444692611694336,
792
+ 45.758731842041016,
793
+ 35.64218521118164,
794
+ 25.07830238342285,
795
+ 40.10424041748047,
796
+ 25.981534957885742,
797
+ 22.0308780670166,
798
+ 29.345321655273438,
799
+ 20.556716918945312,
800
+ 51.5158805847168,
801
+ 52.06641387939453,
802
+ 101.09721374511719,
803
+ 52.499820709228516,
804
+ 46.33478546142578,
805
+ 49.514915466308594,
806
+ 58.03617858886719,
807
+ 135.25387573242188,
808
+ 8.463264465332031
809
+ ],
810
+ "std": [
811
+ 90.01871490478516,
812
+ 81.59170532226562,
813
+ 70.57425689697266,
814
+ 72.62435913085938,
815
+ 67.44258880615234,
816
+ 68.53550720214844,
817
+ 59.094417572021484,
818
+ 68.94862365722656,
819
+ 64.9471664428711,
820
+ 64.55023956298828,
821
+ 52.70843505859375,
822
+ 52.58894348144531,
823
+ 49.775901794433594,
824
+ 49.110530853271484,
825
+ 47.18476104736328,
826
+ 46.478973388671875,
827
+ 50.498287200927734,
828
+ 23.897161483764574,
829
+ 51.42272186279297,
830
+ 52.36309814453125,
831
+ 48.493064880371094,
832
+ 44.10708236694336,
833
+ 44.86425018310547,
834
+ 41.64130401611328,
835
+ 43.47319412231445,
836
+ 45.490352630615234,
837
+ 46.24700927734375,
838
+ 38.542579650878906,
839
+ 41.361751556396484,
840
+ 45.945743560791016,
841
+ 33.42849349975586,
842
+ 35.18679428100586,
843
+ 35.95274353027344,
844
+ 36.03355407714844,
845
+ 36.32166290283203,
846
+ 35.692813873291016,
847
+ 36.22328567504883,
848
+ 45.47581481933594,
849
+ 4.484427928924535,
850
+ 37.25047302246094,
851
+ 37.08232879638672,
852
+ 43.08758544921875,
853
+ 39.07864761352539,
854
+ 55.9732551574707,
855
+ 0.49868059158325195
856
+ ],
857
+ "q01": [
858
+ 0.0,
859
+ 0.0,
860
+ 0.0,
861
+ 0.0,
862
+ 0.0,
863
+ 0.0,
864
+ 0.0,
865
+ 0.0,
866
+ 0.0,
867
+ 0.0,
868
+ 0.0,
869
+ 0.0,
870
+ 0.0,
871
+ 0.0,
872
+ 0.0,
873
+ 0.0,
874
+ 0.0,
875
+ 160.0,
876
+ 0.0,
877
+ 0.0,
878
+ 0.0,
879
+ 0.0,
880
+ 0.0,
881
+ 0.0,
882
+ 0.0,
883
+ 0.0,
884
+ 0.0,
885
+ 0.0,
886
+ 0.0,
887
+ 0.0,
888
+ 0.0,
889
+ 0.0,
890
+ 0.0,
891
+ 0.0,
892
+ 0.0,
893
+ 0.0,
894
+ 6.0,
895
+ 0.0,
896
+ 92.0,
897
+ 0.0,
898
+ 0.0,
899
+ 0.0,
900
+ 0.0,
901
+ 10.0,
902
+ 8.0
903
+ ],
904
+ "q99": [
905
+ 255.0,
906
+ 255.0,
907
+ 255.0,
908
+ 255.0,
909
+ 255.0,
910
+ 255.0,
911
+ 202.0,
912
+ 255.0,
913
+ 255.0,
914
+ 255.0,
915
+ 224.0,
916
+ 223.0,
917
+ 216.0,
918
+ 226.0,
919
+ 189.0,
920
+ 182.0,
921
+ 221.0,
922
+ 255.0,
923
+ 201.0,
924
+ 205.0,
925
+ 195.0,
926
+ 156.0,
927
+ 174.0,
928
+ 151.0,
929
+ 136.0,
930
+ 165.0,
931
+ 159.0,
932
+ 161.0,
933
+ 177.0,
934
+ 190.0,
935
+ 117.0,
936
+ 146.81999999999243,
937
+ 150.0,
938
+ 117.0,
939
+ 139.0,
940
+ 135.0,
941
+ 139.0,
942
+ 197.0,
943
+ 105.0,
944
+ 150.0,
945
+ 149.0,
946
+ 178.0,
947
+ 155.0,
948
+ 246.0,
949
+ 9.0
950
+ ]
951
+ }
952
+ },
953
+ "action": {
954
+ "left_arm": {
955
+ "min": [
956
+ -0.13345633447170258,
957
+ 0.0,
958
+ -0.1426602154970169,
959
+ -2.647650957107544,
960
+ -0.2070874124765396,
961
+ 0.8743690252304077,
962
+ -0.1196235716342926
963
+ ],
964
+ "max": [
965
+ 0.08897088468074799,
966
+ 1.1984225238848012e-05,
967
+ -0.05215534567832947,
968
+ -2.57401967048645,
969
+ -0.041417479515075684,
970
+ 1.0929492712020874,
971
+ -0.00920388475060463
972
+ ],
973
+ "mean": [
974
+ -0.05460730567574501,
975
+ 4.824631805178115e-10,
976
+ -0.10622044652700424,
977
+ -2.6099092960357666,
978
+ -0.19246388971805573,
979
+ 1.0605239868164062,
980
+ -0.1107916384935379
981
+ ],
982
+ "std": [
983
+ 0.05640007182955742,
984
+ 7.603659923915984e-08,
985
+ 0.01339804008603096,
986
+ 0.016931086778640747,
987
+ 0.032859206199645996,
988
+ 0.07246595621109009,
989
+ 0.028103046119213104
990
+ ],
991
+ "q01": [
992
+ -0.13345633447170258,
993
+ 0.0,
994
+ -0.1426602154970169,
995
+ -2.647650957107544,
996
+ -0.2070874124765396,
997
+ 0.8743690252304077,
998
+ -0.1196235716342926
999
+ ],
1000
+ "q99": [
1001
+ 0.08897088468074799,
1002
+ 0.0,
1003
+ -0.05215534567832947,
1004
+ -2.57401967048645,
1005
+ -0.041417479515075684,
1006
+ 1.0929492712020874,
1007
+ -0.00920388475060463
1008
+ ]
1009
+ },
1010
+ "right_arm": {
1011
+ "min": [
1012
+ -0.5568350553512573,
1013
+ -0.6979612708091736,
1014
+ -0.598252534866333,
1015
+ -2.916097402572632,
1016
+ -0.6419709324836731,
1017
+ -1.1942040920257568,
1018
+ -0.5614369511604309
1019
+ ],
1020
+ "max": [
1021
+ 0.37582528591156006,
1022
+ 0.0,
1023
+ 0.5537670850753784,
1024
+ -1.2348545789718628,
1025
+ 1.9420197010040283,
1026
+ 1.5700000524520874,
1027
+ 1.3920875787734985
1028
+ ],
1029
+ "mean": [
1030
+ -0.12528294324874878,
1031
+ -0.07712691277265549,
1032
+ 0.025691134855151176,
1033
+ -2.2697131633758545,
1034
+ 0.6326426863670349,
1035
+ 0.6641106009483337,
1036
+ 0.40074852108955383
1037
+ ],
1038
+ "std": [
1039
+ 0.14245375990867615,
1040
+ 0.10507778078317642,
1041
+ 0.1611136645078659,
1042
+ 0.3662443161010742,
1043
+ 0.5646618604660034,
1044
+ 0.5111078023910522,
1045
+ 0.3725387156009674
1046
+ ],
1047
+ "q01": [
1048
+ -0.48139384508132926,
1049
+ -0.5568350553512573,
1050
+ -0.34207770228385925,
1051
+ -2.759631395339966,
1052
+ -0.3520486056804657,
1053
+ -0.9341943264007568,
1054
+ -0.36125248670578003
1055
+ ],
1056
+ "q99": [
1057
+ 0.19844446450471856,
1058
+ 0.0,
1059
+ 0.3850291669368744,
1060
+ -1.4220001697540283,
1061
+ 1.6636021137237549,
1062
+ 1.3690778017044067,
1063
+ 1.1320778131484985
1064
+ ]
1065
+ },
1066
+ "left_hand": {
1067
+ "min": [
1068
+ -0.7731263041496277,
1069
+ -3.1461946964263916,
1070
+ -3.1247189044952393,
1071
+ -0.251572847366333,
1072
+ -0.1871456503868103,
1073
+ 1.1382137537002563,
1074
+ 0.5875146389007568,
1075
+ 0.7439807057380676,
1076
+ -0.10124272853136063,
1077
+ 1.274738073348999,
1078
+ 0.5737088322639465,
1079
+ 0.6442719101905823,
1080
+ 0.01073786523193121,
1081
+ 1.1550875902175903,
1082
+ 0.6059224009513855,
1083
+ 0.7133010625839233,
1084
+ -0.11198059469461441,
1085
+ 0.977145791053772,
1086
+ 0.5660389065742493,
1087
+ 0.849825382232666
1088
+ ],
1089
+ "max": [
1090
+ 0.09357282519340515,
1091
+ 1.572330355644226,
1092
+ 0.5154175162315369,
1093
+ 38.262081146240234,
1094
+ -0.15493206679821014,
1095
+ 1.2701361179351807,
1096
+ 0.6442719101905823,
1097
+ 0.7593204975128174,
1098
+ -0.07209710031747818,
1099
+ 1.3545050621032715,
1100
+ 0.6135923266410828,
1101
+ 0.6810874938964844,
1102
+ 0.04448544234037399,
1103
+ 1.2394565343856812,
1104
+ 0.6458058953285217,
1105
+ 0.7332428097724915,
1106
+ 0.0951068103313446,
1107
+ 1.0753204822540283,
1108
+ 0.630466103553772,
1109
+ 0.8774369955062866
1110
+ ],
1111
+ "mean": [
1112
+ -0.6870681643486023,
1113
+ 1.508124828338623,
1114
+ -0.7994419932365417,
1115
+ -0.21659596264362335,
1116
+ -0.16863712668418884,
1117
+ 1.1971112489700317,
1118
+ 0.6194714307785034,
1119
+ 0.7555537223815918,
1120
+ -0.08983089029788971,
1121
+ 1.3010036945343018,
1122
+ 0.5898141860961914,
1123
+ 0.6669419407844543,
1124
+ 0.02537773735821247,
1125
+ 1.1902155876159668,
1126
+ 0.6249393224716187,
1127
+ 0.7263370752334595,
1128
+ 0.022195693105459213,
1129
+ 1.0093179941177368,
1130
+ 0.6186497807502747,
1131
+ 0.860866904258728
1132
+ ],
1133
+ "std": [
1134
+ 0.045237179845571365,
1135
+ 0.03564992547035217,
1136
+ 0.6161414980888367,
1137
+ 0.2437615543603897,
1138
+ 0.009025044739246368,
1139
+ 0.034271053969859265,
1140
+ 0.011653272435068935,
1141
+ 0.002153520006682371,
1142
+ 0.007936293259263039,
1143
+ 0.026009514927864075,
1144
+ 0.012295167893170747,
1145
+ 0.011742322705686056,
1146
+ 0.008401543833315372,
1147
+ 0.02993856929242466,
1148
+ 0.01184605155140275,
1149
+ 0.006560774054376007,
1150
+ 0.06808628141880035,
1151
+ 0.02841401472687697,
1152
+ 0.005363619886339923,
1153
+ 0.008137290365992644
1154
+ ],
1155
+ "q01": [
1156
+ -0.7731263041496277,
1157
+ 1.492563247680664,
1158
+ -1.162757396697998,
1159
+ -0.25003886222839355,
1160
+ -0.18561168015003204,
1161
+ 1.1596894264221191,
1162
+ 0.5997864603996277,
1163
+ 0.7470486164093018,
1164
+ -0.09970875084400177,
1165
+ 1.2762720584869385,
1166
+ 0.5737088322639465,
1167
+ 0.6442719101905823,
1168
+ 0.013805827125906944,
1169
+ 1.1596894264221191,
1170
+ 0.6105243563652039,
1171
+ 0.7133010625839233,
1172
+ -0.11198059469461441,
1173
+ 0.983281672000885,
1174
+ 0.6059224009513855,
1175
+ 0.849825382232666
1176
+ ],
1177
+ "q99": [
1178
+ -0.6243301630020142,
1179
+ 1.5692623853683472,
1180
+ 0.5138835906982422,
1181
+ -0.20248545706272125,
1182
+ -0.15493206679821014,
1183
+ 1.2701361179351807,
1184
+ 0.6442719101905823,
1185
+ 0.7593204975128174,
1186
+ -0.07209710031747818,
1187
+ 1.3514370918273926,
1188
+ 0.6105243563652039,
1189
+ 0.676485538482666,
1190
+ 0.04448544234037399,
1191
+ 1.2379225492477417,
1192
+ 0.6442719101905823,
1193
+ 0.731708824634552,
1194
+ 0.0951068103313446,
1195
+ 1.06918466091156,
1196
+ 0.6258641481399536,
1197
+ 0.8774369955062866
1198
+ ]
1199
+ },
1200
+ "right_hand": {
1201
+ "min": [
1202
+ 0.041871316730976105,
1203
+ -1.7027186155319214,
1204
+ -0.4724660813808441,
1205
+ 0.01439326535910368,
1206
+ 0.009813589975237846,
1207
+ 0.07850871980190277,
1208
+ 0.041217077523469925,
1209
+ 0.05037643015384674,
1210
+ 0.00327119673602283,
1211
+ 0.08570535480976105,
1212
+ 0.039908599108457565,
1213
+ 0.04448827728629112,
1214
+ -0.05000000074505806,
1215
+ 0.07981719821691513,
1216
+ 0.041871316730976105,
1217
+ 0.0484137125313282,
1218
+ -0.15000000596046448,
1219
+ 0.06411545723676682,
1220
+ 0.039908599108457565,
1221
+ 0.05757306143641472
1222
+ ],
1223
+ "max": [
1224
+ 0.7685243487358093,
1225
+ -1.477223515510559,
1226
+ 1.1121361255645752,
1227
+ 0.2316311001777649,
1228
+ 0.17947575449943542,
1229
+ 1.2000000476837158,
1230
+ 0.6299999952316284,
1231
+ 0.7699999809265137,
1232
+ 0.1089126393198967,
1233
+ 1.309999942779541,
1234
+ 0.6100000143051147,
1235
+ 0.6800000071525574,
1236
+ 0.00920388475060463,
1237
+ 1.228718638420105,
1238
+ 0.6399999856948853,
1239
+ 0.7400000095367432,
1240
+ -0.009813589975237846,
1241
+ 0.9955535531044006,
1242
+ 0.6181942820549011,
1243
+ 0.8927768468856812
1244
+ ],
1245
+ "mean": [
1246
+ 0.3915060758590698,
1247
+ -1.5039616823196411,
1248
+ 0.6489821672439575,
1249
+ 0.1309857815504074,
1250
+ 0.09314118325710297,
1251
+ 0.7194306254386902,
1252
+ 0.3784645199775696,
1253
+ 0.4631229341030121,
1254
+ 0.036047715693712234,
1255
+ 0.7880135178565979,
1256
+ 0.36028262972831726,
1257
+ 0.4067154824733734,
1258
+ -0.025767970830202103,
1259
+ 0.7310779094696045,
1260
+ 0.3812447190284729,
1261
+ 0.44381776452064514,
1262
+ -0.08370734006166458,
1263
+ 0.5908436179161072,
1264
+ 0.36783817410469055,
1265
+ 0.5302816033363342
1266
+ ],
1267
+ "std": [
1268
+ 0.21117232739925385,
1269
+ 0.018322037532927025,
1270
+ 0.36187946796417236,
1271
+ 0.07061396539211273,
1272
+ 0.05121612548828125,
1273
+ 0.3801894783973694,
1274
+ 0.2004339098930359,
1275
+ 0.2451576292514801,
1276
+ 0.02603093534708023,
1277
+ 0.4172958433628082,
1278
+ 0.18836598098278046,
1279
+ 0.21421349048614502,
1280
+ 0.015439133159816265,
1281
+ 0.38597550988197327,
1282
+ 0.20032943785190582,
1283
+ 0.23426246643066406,
1284
+ 0.04407105594873428,
1285
+ 0.31342390179634094,
1286
+ 0.19525128602981567,
1287
+ 0.2810407280921936
1288
+ ],
1289
+ "q01": [
1290
+ 0.09406757354736328,
1291
+ -1.604543924331665,
1292
+ 0.14311747252941132,
1293
+ 0.030479609966278076,
1294
+ 0.02204708755016327,
1295
+ 0.17637670040130615,
1296
+ 0.09259776771068573,
1297
+ 0.11317504942417145,
1298
+ 0.007349029183387756,
1299
+ 0.19254456460475922,
1300
+ 0.08965815603733063,
1301
+ 0.09994679689407349,
1302
+ -0.05000000074505806,
1303
+ 0.17931631207466125,
1304
+ 0.09406757354736328,
1305
+ 0.1087656319141388,
1306
+ -0.15000000596046448,
1307
+ 0.14404097199440002,
1308
+ 0.08965815603733063,
1309
+ 0.1293429136276245
1310
+ ],
1311
+ "q99": [
1312
+ 0.6795535087585449,
1313
+ -1.5,
1314
+ 1.104466199874878,
1315
+ 0.22242721915245056,
1316
+ 0.17640778422355652,
1317
+ 1.2000000476837158,
1318
+ 0.6299999952316284,
1319
+ 0.7699999809265137,
1320
+ 0.10124272853136063,
1321
+ 1.309999942779541,
1322
+ 0.6100000143051147,
1323
+ 0.6800000071525574,
1324
+ -0.006927184294909239,
1325
+ 1.2200000286102295,
1326
+ 0.6399999856948853,
1327
+ 0.7400000095367432,
1328
+ -0.02204708755016327,
1329
+ 0.9800000190734863,
1330
+ 0.6105243563652039,
1331
+ 0.8799999952316284
1332
+ ]
1333
+ }
1334
+ },
1335
+ "relative_action": {}
1336
+ }
1337
+ }
experiment_cfg/final_model_config.json ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ "letter_box_transform": false,
17
+ "extra_augmentation_config": null,
18
+ "attn_supervision_layers": null,
19
+ "vision_kl_only": false,
20
+ "attn_supervision_weight": 0.5,
21
+ "attn_supervision_loss": "kl",
22
+ "attn_supervision_margin": 0.9,
23
+ "apply_sincos_state_encoding": false,
24
+ "use_percentiles": true,
25
+ "use_relative_action": false,
26
+ "max_state_dim": 144,
27
+ "max_action_dim": 132,
28
+ "action_horizon": 40,
29
+ "hidden_size": 1024,
30
+ "input_embedding_dim": 1536,
31
+ "state_history_length": 1,
32
+ "add_pos_embed": true,
33
+ "attn_dropout": 0.2,
34
+ "use_vlln": true,
35
+ "max_seq_len": 1024,
36
+ "use_alternate_vl_dit": true,
37
+ "attend_text_every_n_blocks": 2,
38
+ "diffusion_model_cfg": {
39
+ "attention_head_dim": 48,
40
+ "dropout": 0.2,
41
+ "final_dropout": true,
42
+ "interleave_self_attention": true,
43
+ "norm_type": "ada_norm",
44
+ "num_attention_heads": 32,
45
+ "num_layers": 32,
46
+ "output_dim": 1024,
47
+ "positional_embeddings": null
48
+ },
49
+ "num_inference_timesteps": 4,
50
+ "noise_beta_alpha": 1.5,
51
+ "noise_beta_beta": 1.0,
52
+ "noise_s": 0.999,
53
+ "num_timestep_buckets": 1000,
54
+ "ttrtc_enabled": false,
55
+ "ttrtc_max_delay_steps": 10,
56
+ "ttrtc_zero_delay_probability": 0.1,
57
+ "ttrtc_policy_hz": 15.0,
58
+ "tune_projector": true,
59
+ "tune_diffusion_model": true,
60
+ "tune_vlln": true,
61
+ "state_dropout_prob": 0.2,
62
+ "exclude_state": false,
63
+ "use_mean_std": false,
64
+ "max_num_embodiments": 32
65
+ }
experiment_cfg/final_processor_config.json ADDED
The diff for this file is too large to render. See raw diff
 
model-00001-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:42460e03ffba1c8026f1e35a2481224b11349cb5dd512508295a8e3a0e658255
3
+ size 4986649584
model-00002-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41c7c929cab21bf7259817b0d297631bd4ff15398146bc774ea446f22db31785
3
+ size 4970792616
model-00003-of-00003.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0204ecca70aaeaeffe52c8900c382fcced761902ca35491195daace8c26d9ed3
3
+ size 2620331560
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a402027af16303ec233ecae6216b1781137328d3e58777c7e32af908e2222a6
3
+ size 12967740438
processor_config.json ADDED
@@ -0,0 +1,1129 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ ],
1020
+ "sin_cos_embedding_keys": null,
1021
+ "mean_std_embedding_keys": null,
1022
+ "action_configs": null
1023
+ },
1024
+ "state": {
1025
+ "delta_indices": [
1026
+ 0
1027
+ ],
1028
+ "modality_keys": [
1029
+ "left_arm",
1030
+ "right_arm",
1031
+ "left_hand",
1032
+ "right_hand",
1033
+ "tactile_left",
1034
+ "tactile_right"
1035
+ ],
1036
+ "sin_cos_embedding_keys": null,
1037
+ "mean_std_embedding_keys": null,
1038
+ "action_configs": null
1039
+ },
1040
+ "action": {
1041
+ "delta_indices": [
1042
+ 0,
1043
+ 1,
1044
+ 2,
1045
+ 3,
1046
+ 4,
1047
+ 5,
1048
+ 6,
1049
+ 7,
1050
+ 8,
1051
+ 9,
1052
+ 10,
1053
+ 11,
1054
+ 12,
1055
+ 13,
1056
+ 14,
1057
+ 15
1058
+ ],
1059
+ "modality_keys": [
1060
+ "left_arm",
1061
+ "right_arm",
1062
+ "left_hand",
1063
+ "right_hand"
1064
+ ],
1065
+ "sin_cos_embedding_keys": null,
1066
+ "mean_std_embedding_keys": null,
1067
+ "action_configs": [
1068
+ {
1069
+ "rep": "ABSOLUTE",
1070
+ "type": "NON_EEF",
1071
+ "format": "DEFAULT",
1072
+ "state_key": null
1073
+ },
1074
+ {
1075
+ "rep": "ABSOLUTE",
1076
+ "type": "NON_EEF",
1077
+ "format": "DEFAULT",
1078
+ "state_key": null
1079
+ },
1080
+ {
1081
+ "rep": "ABSOLUTE",
1082
+ "type": "NON_EEF",
1083
+ "format": "DEFAULT",
1084
+ "state_key": null
1085
+ },
1086
+ {
1087
+ "rep": "ABSOLUTE",
1088
+ "type": "NON_EEF",
1089
+ "format": "DEFAULT",
1090
+ "state_key": null
1091
+ }
1092
+ ]
1093
+ },
1094
+ "language": {
1095
+ "delta_indices": [
1096
+ 0
1097
+ ],
1098
+ "modality_keys": [
1099
+ "task"
1100
+ ],
1101
+ "sin_cos_embedding_keys": null,
1102
+ "mean_std_embedding_keys": null,
1103
+ "action_configs": null
1104
+ }
1105
+ }
1106
+ },
1107
+ "image_crop_size": null,
1108
+ "image_target_size": null,
1109
+ "use_albumentations": true,
1110
+ "random_rotation_angle": null,
1111
+ "color_jitter_params": null,
1112
+ "shortest_image_edge": 256,
1113
+ "crop_fraction": 0.9,
1114
+ "letter_box_transform": true,
1115
+ "model_name": "nvidia/Cosmos-Reason2-2B",
1116
+ "model_type": "qwen",
1117
+ "formalize_language": true,
1118
+ "max_state_dim": 144,
1119
+ "max_action_dim": 132,
1120
+ "max_action_horizon": 40,
1121
+ "use_percentiles": true,
1122
+ "use_mean_std": false,
1123
+ "clip_outliers": true,
1124
+ "apply_sincos_state_encoding": false,
1125
+ "use_relative_action": true,
1126
+ "exclude_state": false,
1127
+ "state_dropout_prob": 0.2
1128
+ }
1129
+ }
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6584a122167f2280ebe505b20867fe908c89b7a4063dfdf2e12522a90c6c301f
3
+ size 14645
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1cfac7dd5c42f227b537b0eeec342b62379e7bf2b2ed74e2dfbb205073f3a9d1
3
+ size 1465
statistics.json ADDED
The diff for this file is too large to render. See raw diff
 
trainer_state.json ADDED
The diff for this file is too large to render. See raw diff
 
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa2377a24bef7d7e6ba849a63d452940ffc05a70d4d6b6027898af48072348d2
3
+ size 5969
wandb_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"project": "finetune-gr00t-n1d7", "run_id": "802_v21_frozen_vlm_action_100k_20261104"}