Dongkkka commited on
Commit
1dc0f98
·
verified ·
1 Parent(s): a323f83

Add files using upload-large-folder tool

Browse files
config.json ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "action_horizon": 50,
3
+ "add_pos_embed": true,
4
+ "apply_sincos_state_encoding": true,
5
+ "architectures": [
6
+ "Gr00tN1d6"
7
+ ],
8
+ "attn_dropout": 0.2,
9
+ "attn_implementation": null,
10
+ "backbone_embedding_dim": 2048,
11
+ "backbone_model_type": "eagle",
12
+ "backbone_trainable_params_fp32": true,
13
+ "collator_overwrite_image_inputs": false,
14
+ "color_jitter_params": {
15
+ "brightness": 0.1,
16
+ "contrast": 0.1,
17
+ "hue": 0.1,
18
+ "saturation": 0.1
19
+ },
20
+ "crop_fraction": 0.95,
21
+ "diffusion_model_cfg": {
22
+ "attention_head_dim": 48,
23
+ "dropout": 0.2,
24
+ "final_dropout": true,
25
+ "interleave_self_attention": true,
26
+ "norm_type": "ada_norm",
27
+ "num_attention_heads": 32,
28
+ "num_layers": 32,
29
+ "output_dim": 1024,
30
+ "positional_embeddings": null
31
+ },
32
+ "eagle_collator": true,
33
+ "formalize_language": true,
34
+ "gemma_collator": false,
35
+ "hidden_size": 1024,
36
+ "image_crop_size": null,
37
+ "image_target_size": null,
38
+ "input_embedding_dim": 1536,
39
+ "load_bf16": true,
40
+ "max_action_dim": 128,
41
+ "max_num_embodiments": 32,
42
+ "max_seq_len": 1024,
43
+ "max_state_dim": 128,
44
+ "model_dtype": "bfloat16",
45
+ "model_name": "nvidia/Eagle-Block2A-2B-v2",
46
+ "model_type": "Gr00tN1d6",
47
+ "noise_beta_alpha": 1.5,
48
+ "noise_beta_beta": 1.0,
49
+ "noise_s": 0.999,
50
+ "num_inference_timesteps": 4,
51
+ "num_timestep_buckets": 1000,
52
+ "random_rotation_angle": null,
53
+ "reproject_vision": false,
54
+ "select_layer": 16,
55
+ "shortest_image_edge": 256,
56
+ "state_dropout_prob": 0.0,
57
+ "torch_dtype": "bfloat16",
58
+ "transformers_version": "4.51.3",
59
+ "tune_diffusion_model": true,
60
+ "tune_llm": false,
61
+ "tune_projector": true,
62
+ "tune_top_llm_layers": 4,
63
+ "tune_visual": false,
64
+ "tune_vlln": true,
65
+ "use_albumentations_transforms": true,
66
+ "use_alternate_vl_dit": true,
67
+ "use_flash_attention": true,
68
+ "use_relative_action": true,
69
+ "use_vlln": true
70
+ }
embodiment_id.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "robocasa_panda_omron": 13,
3
+ "gr1": 20,
4
+ "behavior_r1_pro": 24,
5
+ "unitree_g1": 8,
6
+ "oxe_google": 0,
7
+ "oxe_widowx": 1,
8
+ "libero_panda": 2,
9
+ "oxe_droid": 16,
10
+ "new_embodiment": 10
11
+ }
experiment_cfg/conf.yaml ADDED
@@ -0,0 +1,208 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ load_config_path: null
2
+ model:
3
+ model_type: Gr00tN1d6
4
+ model_dtype: bfloat16
5
+ model_name: nvidia/Eagle-Block2A-2B-v2
6
+ backbone_model_type: eagle
7
+ model_revision: null
8
+ tune_top_llm_layers: 4
9
+ backbone_embedding_dim: 2048
10
+ tune_llm: false
11
+ tune_visual: false
12
+ select_layer: 16
13
+ reproject_vision: false
14
+ use_flash_attention: true
15
+ load_bf16: false
16
+ collator_overwrite_image_inputs: false
17
+ eagle_collator: true
18
+ backbone_trainable_params_fp32: true
19
+ image_crop_size: null
20
+ image_target_size: null
21
+ shortest_image_edge: 256
22
+ crop_fraction: 0.95
23
+ random_rotation_angle: null
24
+ color_jitter_params:
25
+ brightness: 0.3
26
+ contrast: 0.4
27
+ saturation: 0.5
28
+ hue: 0.08
29
+ use_albumentations_transforms: true
30
+ formalize_language: true
31
+ apply_sincos_state_encoding: false
32
+ use_relative_action: true
33
+ max_state_dim: 29
34
+ max_action_dim: 29
35
+ action_horizon: 16
36
+ hidden_size: 1024
37
+ input_embedding_dim: 1536
38
+ add_pos_embed: true
39
+ attn_dropout: 0.2
40
+ use_vlln: true
41
+ max_seq_len: 1024
42
+ use_alternate_vl_dit: true
43
+ attend_text_every_n_blocks: 2
44
+ diffusion_model_cfg:
45
+ positional_embeddings: null
46
+ num_layers: 32
47
+ num_attention_heads: 32
48
+ attention_head_dim: 48
49
+ norm_type: ada_norm
50
+ dropout: 0.2
51
+ final_dropout: true
52
+ output_dim: 1024
53
+ interleave_self_attention: true
54
+ num_inference_timesteps: 4
55
+ noise_beta_alpha: 1.5
56
+ noise_beta_beta: 1.0
57
+ noise_s: 0.999
58
+ num_timestep_buckets: 1000
59
+ tune_projector: true
60
+ tune_diffusion_model: true
61
+ tune_vlln: true
62
+ state_dropout_prob: 0.0
63
+ state_additive_noise_scale: 0.0
64
+ max_num_embodiments: 32
65
+ data:
66
+ datasets:
67
+ - dataset_paths:
68
+ - /workspace/checkpoints/Dongkkka/Box_Packing_0320_lerobot
69
+ embodiment_tag: new_embodiment
70
+ mix_ratio: 1.0
71
+ dataset_type: physical_embodiment
72
+ val_dataset_path: null
73
+ modality_configs:
74
+ new_embodiment:
75
+ video:
76
+ delta_indices:
77
+ - 0
78
+ modality_keys:
79
+ - cam_left_head
80
+ sin_cos_embedding_keys: null
81
+ mean_std_embedding_keys: null
82
+ action_configs: null
83
+ state:
84
+ delta_indices:
85
+ - 0
86
+ modality_keys:
87
+ - arm_left
88
+ - arm_right
89
+ sin_cos_embedding_keys: null
90
+ mean_std_embedding_keys: null
91
+ action_configs: null
92
+ action:
93
+ delta_indices:
94
+ - 0
95
+ - 1
96
+ - 2
97
+ - 3
98
+ - 4
99
+ - 5
100
+ - 6
101
+ - 7
102
+ - 8
103
+ - 9
104
+ - 10
105
+ - 11
106
+ - 12
107
+ - 13
108
+ - 14
109
+ - 15
110
+ modality_keys:
111
+ - arm_left
112
+ - arm_right
113
+ sin_cos_embedding_keys: null
114
+ mean_std_embedding_keys: null
115
+ action_configs:
116
+ - rep: ABSOLUTE
117
+ type: NON_EEF
118
+ format: DEFAULT
119
+ state_key: null
120
+ - rep: ABSOLUTE
121
+ type: NON_EEF
122
+ format: DEFAULT
123
+ state_key: null
124
+ language:
125
+ delta_indices:
126
+ - 0
127
+ modality_keys:
128
+ - annotation.human.task_description
129
+ sin_cos_embedding_keys: null
130
+ mean_std_embedding_keys: null
131
+ action_configs: null
132
+ download_cache: false
133
+ shard_size: 1024
134
+ episode_sampling_rate: 0.1
135
+ num_shards_per_epoch: 100000
136
+ override_pretraining_statistics: false
137
+ mode: single_turn
138
+ random_chop: 0.0
139
+ mock_dataset_mode: false
140
+ shuffle: true
141
+ seed: 42
142
+ multiprocessing_context: fork
143
+ allow_padding: false
144
+ subsample_ratio: 1.0
145
+ image_crop_size:
146
+ - 244
147
+ - 244
148
+ image_target_size:
149
+ - 224
150
+ - 224
151
+ video_backend: torchcodec
152
+ training:
153
+ output_dir: /workspace/checkpoints/model/Box_Packing_0320_lerobot_model
154
+ experiment_name: null
155
+ max_steps: 50000
156
+ global_batch_size: 48
157
+ batch_size: null
158
+ gradient_accumulation_steps: 1
159
+ learning_rate: 0.0001
160
+ lr_scheduler_type: cosine
161
+ weight_decay: 1.0e-05
162
+ warmup_ratio: 0.05
163
+ warmup_steps: 0
164
+ max_grad_norm: 1.0
165
+ optim: adamw_torch
166
+ start_from_checkpoint: nvidia/GR00T-N1.6-3B
167
+ tf32: true
168
+ fp16: false
169
+ bf16: true
170
+ eval_bf16: true
171
+ logging_steps: 10
172
+ save_steps: 10000
173
+ save_total_limit: 10
174
+ save_vl_model: false
175
+ upload_checkpoints: false
176
+ upload_every: 1000
177
+ upload_last_n_checkpoints: 5
178
+ max_concurrent_uploads: 2
179
+ eval_strategy: 'no'
180
+ eval_steps: 500
181
+ eval_set_split_ratio: 0.1
182
+ eval_batch_size: 2
183
+ save_best_eval_metric_name: ''
184
+ save_best_eval_metric_greater_is_better: true
185
+ deepspeed_stage: 2
186
+ gradient_checkpointing: false
187
+ transformers_trust_remote_code: true
188
+ transformers_local_files_only: false
189
+ transformers_cache_dir: null
190
+ transformers_access_token: null
191
+ use_ddp: false
192
+ ddp_bucket_cap_mb: 100
193
+ num_gpus: 1
194
+ dataloader_num_workers: 8
195
+ remove_unused_columns: false
196
+ use_wandb: false
197
+ wandb_project: finetune-gr00t-n1d6
198
+ enable_profiling: false
199
+ max_retries: 3
200
+ assert_loss_less_than: null
201
+ add_rl_callback: false
202
+ enable_open_loop_eval: false
203
+ open_loop_eval_traj_ids:
204
+ - 0
205
+ open_loop_eval_steps_per_traj: 100
206
+ open_loop_eval_plot_indices: null
207
+ max_steps: 50000
208
+ save_steps: 10000
experiment_cfg/config.yaml ADDED
@@ -0,0 +1,241 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ - /workspace/checkpoints/Dongkkka/Box_Packing_0320_lerobot
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
+ delta_indices:
39
+ - 0
40
+ - 1
41
+ - 2
42
+ - 3
43
+ - 4
44
+ - 5
45
+ - 6
46
+ - 7
47
+ - 8
48
+ - 9
49
+ - 10
50
+ - 11
51
+ - 12
52
+ - 13
53
+ - 14
54
+ - 15
55
+ mean_std_embedding_keys: null
56
+ modality_keys:
57
+ - arm_left
58
+ - arm_right
59
+ sin_cos_embedding_keys: null
60
+ language: !!python/object:gr00t.data.types.ModalityConfig
61
+ action_configs: null
62
+ delta_indices:
63
+ - 0
64
+ mean_std_embedding_keys: null
65
+ modality_keys:
66
+ - annotation.human.task_description
67
+ sin_cos_embedding_keys: null
68
+ state: !!python/object:gr00t.data.types.ModalityConfig
69
+ action_configs: null
70
+ delta_indices:
71
+ - 0
72
+ mean_std_embedding_keys: null
73
+ modality_keys:
74
+ - arm_left
75
+ - arm_right
76
+ sin_cos_embedding_keys: null
77
+ video: !!python/object:gr00t.data.types.ModalityConfig
78
+ action_configs: null
79
+ delta_indices:
80
+ - 0
81
+ mean_std_embedding_keys: null
82
+ modality_keys:
83
+ - cam_left_head
84
+ sin_cos_embedding_keys: null
85
+ mode: single_turn
86
+ multiprocessing_context: fork
87
+ num_shards_per_epoch: 100000
88
+ override_pretraining_statistics: false
89
+ random_chop: 0.0
90
+ seed: 42
91
+ shard_size: 1024
92
+ shuffle: true
93
+ subsample_ratio: 1.0
94
+ video_backend: torchcodec
95
+ load_config_path: null
96
+ model: !!python/object:gr00t.configs.model.gr00t_n1d6.Gr00tN1d6Config
97
+ _attn_implementation_autoset: false
98
+ _attn_implementation_internal: null
99
+ _commit_hash: null
100
+ _name_or_path: ''
101
+ add_cross_attention: false
102
+ architectures: null
103
+ backbone_model_type: eagle
104
+ backbone_trainable_params_fp32: true
105
+ bad_words_ids: null
106
+ begin_suppress_tokens: null
107
+ bos_token_id: null
108
+ chunk_size_feed_forward: 0
109
+ color_jitter_params:
110
+ brightness: 0.3
111
+ contrast: 0.4
112
+ hue: 0.08
113
+ saturation: 0.5
114
+ cross_attention_hidden_size: null
115
+ decoder_start_token_id: null
116
+ diffusion_model_cfg:
117
+ attention_head_dim: 48
118
+ dropout: 0.2
119
+ final_dropout: true
120
+ interleave_self_attention: true
121
+ norm_type: ada_norm
122
+ num_attention_heads: 32
123
+ num_layers: 32
124
+ output_dim: 1024
125
+ positional_embeddings: null
126
+ diversity_penalty: 0.0
127
+ do_sample: false
128
+ eagle_collator: true
129
+ early_stopping: false
130
+ encoder_no_repeat_ngram_size: 0
131
+ eos_token_id: null
132
+ exponential_decay_length_penalty: null
133
+ finetuning_task: null
134
+ forced_bos_token_id: null
135
+ forced_eos_token_id: null
136
+ id2label:
137
+ 0: LABEL_0
138
+ 1: LABEL_1
139
+ is_decoder: false
140
+ is_encoder_decoder: false
141
+ label2id:
142
+ LABEL_0: 0
143
+ LABEL_1: 1
144
+ length_penalty: 1.0
145
+ load_bf16: false
146
+ max_length: 20
147
+ min_length: 0
148
+ model_name: nvidia/Eagle-Block2A-2B-v2
149
+ no_repeat_ngram_size: 0
150
+ num_beam_groups: 1
151
+ num_beams: 1
152
+ num_return_sequences: 1
153
+ output_attentions: false
154
+ output_hidden_states: false
155
+ output_scores: false
156
+ pad_token_id: null
157
+ prefix: null
158
+ problem_type: null
159
+ pruned_heads: {}
160
+ random_rotation_angle: null
161
+ remove_invalid_values: false
162
+ repetition_penalty: 1.0
163
+ reproject_vision: false
164
+ return_dict: true
165
+ return_dict_in_generate: false
166
+ sep_token_id: null
167
+ state_dropout_prob: 0.0
168
+ suppress_tokens: null
169
+ task_specific_params: null
170
+ temperature: 1.0
171
+ tf_legacy_loss: false
172
+ tie_encoder_decoder: false
173
+ tie_word_embeddings: true
174
+ tokenizer_class: null
175
+ top_k: 50
176
+ top_p: 1.0
177
+ torch_dtype: null
178
+ torchscript: false
179
+ transformers_version: null
180
+ tune_diffusion_model: true
181
+ tune_llm: false
182
+ tune_projector: true
183
+ tune_visual: false
184
+ typical_p: 1.0
185
+ use_bfloat16: false
186
+ use_relative_action: true
187
+ training: !!python/object:gr00t.configs.training.training_config.TrainingConfig
188
+ add_rl_callback: false
189
+ assert_loss_less_than: null
190
+ batch_size: null
191
+ bf16: true
192
+ dataloader_num_workers: 8
193
+ ddp_bucket_cap_mb: 100
194
+ deepspeed_stage: 2
195
+ enable_open_loop_eval: false
196
+ enable_profiling: false
197
+ eval_batch_size: 2
198
+ eval_bf16: true
199
+ eval_set_split_ratio: 0.1
200
+ eval_steps: 500
201
+ eval_strategy: 'no'
202
+ experiment_name: null
203
+ fp16: false
204
+ global_batch_size: 48
205
+ gradient_accumulation_steps: 1
206
+ gradient_checkpointing: false
207
+ learning_rate: 0.0001
208
+ logging_steps: 10
209
+ lr_scheduler_type: cosine
210
+ max_concurrent_uploads: 2
211
+ max_grad_norm: 1.0
212
+ max_retries: 3
213
+ max_steps: 50000
214
+ num_gpus: 1
215
+ open_loop_eval_plot_indices: null
216
+ open_loop_eval_steps_per_traj: 100
217
+ open_loop_eval_traj_ids:
218
+ - 0
219
+ optim: adamw_torch
220
+ output_dir: /workspace/checkpoints/model/Box_Packing_0320_lerobot_model
221
+ remove_unused_columns: false
222
+ save_best_eval_metric_greater_is_better: true
223
+ save_best_eval_metric_name: ''
224
+ save_steps: 10000
225
+ save_total_limit: 10
226
+ save_vl_model: false
227
+ start_from_checkpoint: nvidia/GR00T-N1.6-3B
228
+ tf32: true
229
+ transformers_access_token: null
230
+ transformers_cache_dir: null
231
+ transformers_local_files_only: false
232
+ transformers_trust_remote_code: true
233
+ upload_checkpoints: false
234
+ upload_every: 1000
235
+ upload_last_n_checkpoints: 5
236
+ use_ddp: false
237
+ use_wandb: false
238
+ wandb_project: finetune-gr00t-n1d6
239
+ warmup_ratio: 0.05
240
+ warmup_steps: 0
241
+ weight_decay: 1.0e-05
experiment_cfg/dataset_statistics.json ADDED
@@ -0,0 +1,257 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "new_embodiment": {
3
+ "state": {
4
+ "arm_left": {
5
+ "min": [
6
+ -1.040446400642395,
7
+ -0.00017976337403524667,
8
+ -1.2547483444213867,
9
+ -2.3187437057495117,
10
+ -1.3268334865570068,
11
+ -0.9030593037605286,
12
+ -1.2854517698287964,
13
+ 0.0
14
+ ],
15
+ "max": [
16
+ 0.9663000106811523,
17
+ 0.7367781400680542,
18
+ 0.5400571227073669,
19
+ -0.7592845559120178,
20
+ 1.5720906257629395,
21
+ 1.5657269954681396,
22
+ 0.2666449248790741,
23
+ 1.0
24
+ ],
25
+ "mean": [
26
+ 0.049500223249197006,
27
+ 0.202315092086792,
28
+ -0.14248345792293549,
29
+ -1.7501721382141113,
30
+ -0.11155878007411957,
31
+ 0.15117862820625305,
32
+ -0.20862539112567902,
33
+ 0.421645849943161
34
+ ],
35
+ "std": [
36
+ 0.48967358469963074,
37
+ 0.13500156998634338,
38
+ 0.3402341604232788,
39
+ 0.24770542979240417,
40
+ 0.5848802328109741,
41
+ 0.48957449197769165,
42
+ 0.24857880175113678,
43
+ 0.38907936215400696
44
+ ],
45
+ "q01": [
46
+ -0.94666987657547,
47
+ 0.0,
48
+ -1.068766963481903,
49
+ -2.1590683579444887,
50
+ -1.1686416864395142,
51
+ -0.5971877157688141,
52
+ -1.0124290406703949,
53
+ 0.11042070388793945
54
+ ],
55
+ "q99": [
56
+ 0.8475315928459172,
57
+ 0.5942887216806414,
58
+ 0.3706948429346091,
59
+ -1.028668290376662,
60
+ 1.1839396059513096,
61
+ 1.3001326322555542,
62
+ 0.15171746462583544,
63
+ 1.0
64
+ ]
65
+ },
66
+ "arm_right": {
67
+ "min": [
68
+ -0.6212142705917358,
69
+ -0.8272590637207031,
70
+ -1.2627418041229248,
71
+ -2.7542264461517334,
72
+ -0.49333062767982483,
73
+ -0.7774885892868042,
74
+ -0.7931810617446899,
75
+ 0.0
76
+ ],
77
+ "max": [
78
+ 0.6666824221611023,
79
+ -0.0020013656467199326,
80
+ 0.4161761701107025,
81
+ -0.48478585481643677,
82
+ 1.6210821866989136,
83
+ 0.4454296827316284,
84
+ 0.22090955078601837,
85
+ 0.22853723168373108
86
+ ],
87
+ "mean": [
88
+ -0.12877272069454193,
89
+ -0.2153315395116806,
90
+ -0.04238507151603699,
91
+ -1.434108018875122,
92
+ 0.27934524416923523,
93
+ -0.3126658797264099,
94
+ -0.25879907608032227,
95
+ 0.16445133090019226
96
+ ],
97
+ "std": [
98
+ 0.32200026512145996,
99
+ 0.09513258934020996,
100
+ 0.3042828142642975,
101
+ 0.8548151850700378,
102
+ 0.597506046295166,
103
+ 0.2752605080604553,
104
+ 0.2458854615688324,
105
+ 0.023125283420085907
106
+ ],
107
+ "q01": [
108
+ -0.541259127855301,
109
+ -0.49224483519792556,
110
+ -0.8095595329999924,
111
+ -2.591925394535065,
112
+ -0.35722817331552503,
113
+ -0.7064988255500794,
114
+ -0.6845486432313919,
115
+ 0.12166989594697952
116
+ ],
117
+ "q99": [
118
+ 0.4248299956321717,
119
+ -0.06495329849421971,
120
+ 0.315976083278656,
121
+ -0.5046437382698059,
122
+ 1.4662519514560723,
123
+ 0.2226938612759115,
124
+ 0.20540948137640974,
125
+ 0.22010032832622528
126
+ ]
127
+ }
128
+ },
129
+ "action": {
130
+ "arm_left": {
131
+ "min": [
132
+ -1.0400389432907104,
133
+ -0.3067961633205414,
134
+ -1.2547962665557861,
135
+ -2.322446823120117,
136
+ -1.3268933296203613,
137
+ -0.9087768793106079,
138
+ -1.2854758501052856,
139
+ -0.2644738256931305
140
+ ],
141
+ "max": [
142
+ 0.9664078950881958,
143
+ 0.7378447651863098,
144
+ 0.5445631742477417,
145
+ -0.7593204975128174,
146
+ 1.572330355644226,
147
+ 1.564000129699707,
148
+ 0.2684466242790222,
149
+ 2.259845018386841
150
+ ],
151
+ "mean": [
152
+ 0.051741912961006165,
153
+ 0.1994943916797638,
154
+ -0.14148958027362823,
155
+ -1.7508922815322876,
156
+ -0.11110030859708786,
157
+ 0.14937299489974976,
158
+ -0.20768238604068756,
159
+ 0.7344653010368347
160
+ ],
161
+ "std": [
162
+ 0.48878252506256104,
163
+ 0.14125312864780426,
164
+ 0.3398240804672241,
165
+ 0.24780994653701782,
166
+ 0.5844447016716003,
167
+ 0.4887670874595642,
168
+ 0.24728405475616455,
169
+ 0.967104434967041
170
+ ],
171
+ "q01": [
172
+ -0.9464661478996277,
173
+ -0.11044661700725555,
174
+ -1.0707186460494995,
175
+ -2.1598448753356934,
176
+ -1.1688933372497559,
177
+ -0.5958448052406311,
178
+ -1.0047574043273926,
179
+ 0.0325048454105854
180
+ ],
181
+ "q99": [
182
+ 0.8499020785093352,
183
+ 0.5951845645904541,
184
+ 0.37275734543800354,
185
+ -1.0277671813964844,
186
+ 1.1857671737670898,
187
+ 1.300155520439148,
188
+ 0.15186409652233124,
189
+ 2.249720573425293
190
+ ]
191
+ },
192
+ "arm_right": {
193
+ "min": [
194
+ -0.6212621927261353,
195
+ -0.8283496499061584,
196
+ -1.3606410026550293,
197
+ -2.756563425064087,
198
+ -0.4939418137073517,
199
+ -0.7829904556274414,
200
+ -0.7930680513381958,
201
+ -0.1328582763671875
202
+ ],
203
+ "max": [
204
+ 0.6795535087585449,
205
+ -0.0015339808305725455,
206
+ 0.4218447208404541,
207
+ -0.48473793268203735,
208
+ 1.6244856119155884,
209
+ 0.45033010840415955,
210
+ 0.2208932340145111,
211
+ 0.17424467206001282
212
+ ],
213
+ "mean": [
214
+ -0.1264759600162506,
215
+ -0.21621093153953552,
216
+ -0.0449027381837368,
217
+ -1.4389472007751465,
218
+ 0.2804511487483978,
219
+ -0.31131044030189514,
220
+ -0.25912487506866455,
221
+ 0.09730902314186096
222
+ ],
223
+ "std": [
224
+ 0.3228287994861603,
225
+ 0.09637373685836792,
226
+ 0.30746692419052124,
227
+ 0.8555927872657776,
228
+ 0.5972351431846619,
229
+ 0.27582842111587524,
230
+ 0.2455405592918396,
231
+ 0.02781537175178528
232
+ ],
233
+ "q01": [
234
+ -0.5414952039718628,
235
+ -0.4970097839832306,
236
+ -0.8160777688026428,
237
+ -2.5954954624176025,
238
+ -0.3574175238609314,
239
+ -0.7062913775444031,
240
+ -0.6841554045677185,
241
+ 0.046003878116607666
242
+ ],
243
+ "q99": [
244
+ 0.424912691116333,
245
+ -0.06442718952894211,
246
+ 0.3160000443458557,
247
+ -0.5046796798706055,
248
+ 1.4695535898208618,
249
+ 0.22636891901493073,
250
+ 0.20555342733860016,
251
+ 0.1641203910112381
252
+ ]
253
+ }
254
+ },
255
+ "relative_action": {}
256
+ }
257
+ }
experiment_cfg/final_model_config.json ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "Gr00tN1d6",
3
+ "model_dtype": "bfloat16",
4
+ "model_name": "nvidia/Eagle-Block2A-2B-v2",
5
+ "backbone_model_type": "eagle",
6
+ "model_revision": null,
7
+ "tune_top_llm_layers": 4,
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": true,
15
+ "collator_overwrite_image_inputs": false,
16
+ "eagle_collator": true,
17
+ "backbone_trainable_params_fp32": true,
18
+ "apply_sincos_state_encoding": true,
19
+ "use_relative_action": true,
20
+ "max_state_dim": 128,
21
+ "max_action_dim": 128,
22
+ "action_horizon": 50,
23
+ "hidden_size": 1024,
24
+ "input_embedding_dim": 1536,
25
+ "add_pos_embed": true,
26
+ "attn_dropout": 0.2,
27
+ "use_vlln": true,
28
+ "max_seq_len": 1024,
29
+ "use_alternate_vl_dit": true,
30
+ "attend_text_every_n_blocks": 2,
31
+ "diffusion_model_cfg": {
32
+ "attention_head_dim": 48,
33
+ "dropout": 0.2,
34
+ "final_dropout": true,
35
+ "interleave_self_attention": true,
36
+ "norm_type": "ada_norm",
37
+ "num_attention_heads": 32,
38
+ "num_layers": 32,
39
+ "output_dim": 1024,
40
+ "positional_embeddings": null
41
+ },
42
+ "num_inference_timesteps": 4,
43
+ "noise_beta_alpha": 1.5,
44
+ "noise_beta_beta": 1.0,
45
+ "noise_s": 0.999,
46
+ "num_timestep_buckets": 1000,
47
+ "tune_projector": true,
48
+ "tune_diffusion_model": true,
49
+ "tune_vlln": true,
50
+ "state_dropout_prob": 0.0,
51
+ "state_additive_noise_scale": 0.0,
52
+ "max_num_embodiments": 32
53
+ }
experiment_cfg/final_processor_config.json ADDED
The diff for this file is too large to render. See raw diff
 
model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b756336c9722175b4a898c726985060fd5cb5b6ac316674fc18be50bab14c57e
3
+ size 4990120184
model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ae3c25fd466d72686cf47ef52ebfdf93450d06b32bd2faea2050dcbdc1529964
3
+ size 4823190320
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
processor_config.json ADDED
@@ -0,0 +1,454 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "processor_class": "Gr00tN1d6Processor",
3
+ "processor_kwargs": {
4
+ "modality_configs": {
5
+ "behavior_r1_pro": {
6
+ "video": {
7
+ "delta_indices": [
8
+ 0
9
+ ],
10
+ "modality_keys": [
11
+ "observation.images.rgb.head_256_256",
12
+ "observation.images.rgb.left_wrist_256_256",
13
+ "observation.images.rgb.right_wrist_256_256"
14
+ ],
15
+ "sin_cos_embedding_keys": null,
16
+ "mean_std_embedding_keys": null,
17
+ "action_configs": null
18
+ },
19
+ "state": {
20
+ "delta_indices": [
21
+ 0
22
+ ],
23
+ "modality_keys": [
24
+ "robot_pos",
25
+ "robot_ori_cos",
26
+ "robot_ori_sin",
27
+ "robot_2d_ori",
28
+ "robot_2d_ori_cos",
29
+ "robot_2d_ori_sin",
30
+ "robot_lin_vel",
31
+ "robot_ang_vel",
32
+ "arm_left_qpos",
33
+ "arm_left_qpos_sin",
34
+ "arm_left_qpos_cos",
35
+ "eef_left_pos",
36
+ "eef_left_quat",
37
+ "gripper_left_qpos",
38
+ "arm_right_qpos",
39
+ "arm_right_qpos_sin",
40
+ "arm_right_qpos_cos",
41
+ "eef_right_pos",
42
+ "eef_right_quat",
43
+ "gripper_right_qpos",
44
+ "trunk_qpos"
45
+ ],
46
+ "sin_cos_embedding_keys": null,
47
+ "mean_std_embedding_keys": null,
48
+ "action_configs": null
49
+ },
50
+ "action": {
51
+ "delta_indices": [
52
+ 0,
53
+ 1,
54
+ 2,
55
+ 3,
56
+ 4,
57
+ 5,
58
+ 6,
59
+ 7,
60
+ 8,
61
+ 9,
62
+ 10,
63
+ 11,
64
+ 12,
65
+ 13,
66
+ 14,
67
+ 15,
68
+ 16,
69
+ 17,
70
+ 18,
71
+ 19,
72
+ 20,
73
+ 21,
74
+ 22,
75
+ 23,
76
+ 24,
77
+ 25,
78
+ 26,
79
+ 27,
80
+ 28,
81
+ 29,
82
+ 30,
83
+ 31
84
+ ],
85
+ "modality_keys": [
86
+ "base",
87
+ "torso",
88
+ "left_arm",
89
+ "left_gripper",
90
+ "right_arm",
91
+ "right_gripper"
92
+ ],
93
+ "sin_cos_embedding_keys": null,
94
+ "mean_std_embedding_keys": null,
95
+ "action_configs": [
96
+ {
97
+ "rep": "ABSOLUTE",
98
+ "type": "NON_EEF",
99
+ "format": "DEFAULT",
100
+ "state_key": null
101
+ },
102
+ {
103
+ "rep": "RELATIVE",
104
+ "type": "NON_EEF",
105
+ "format": "DEFAULT",
106
+ "state_key": "trunk_qpos"
107
+ },
108
+ {
109
+ "rep": "RELATIVE",
110
+ "type": "NON_EEF",
111
+ "format": "DEFAULT",
112
+ "state_key": "arm_left_qpos"
113
+ },
114
+ {
115
+ "rep": "ABSOLUTE",
116
+ "type": "NON_EEF",
117
+ "format": "DEFAULT",
118
+ "state_key": null
119
+ },
120
+ {
121
+ "rep": "RELATIVE",
122
+ "type": "NON_EEF",
123
+ "format": "DEFAULT",
124
+ "state_key": "arm_right_qpos"
125
+ },
126
+ {
127
+ "rep": "ABSOLUTE",
128
+ "type": "NON_EEF",
129
+ "format": "DEFAULT",
130
+ "state_key": null
131
+ }
132
+ ]
133
+ },
134
+ "language": {
135
+ "delta_indices": [
136
+ 0
137
+ ],
138
+ "modality_keys": [
139
+ "annotation.human.coarse_action"
140
+ ],
141
+ "sin_cos_embedding_keys": null,
142
+ "mean_std_embedding_keys": null,
143
+ "action_configs": null
144
+ }
145
+ },
146
+ "gr1": {
147
+ "video": {
148
+ "delta_indices": [
149
+ 0
150
+ ],
151
+ "modality_keys": [
152
+ "ego_view_bg_crop_pad_res256_freq20"
153
+ ],
154
+ "sin_cos_embedding_keys": null,
155
+ "mean_std_embedding_keys": null,
156
+ "action_configs": null
157
+ },
158
+ "state": {
159
+ "delta_indices": [
160
+ 0
161
+ ],
162
+ "modality_keys": [
163
+ "left_arm",
164
+ "right_arm",
165
+ "left_hand",
166
+ "right_hand",
167
+ "waist"
168
+ ],
169
+ "sin_cos_embedding_keys": [
170
+ "left_arm",
171
+ "right_arm",
172
+ "left_hand",
173
+ "right_hand",
174
+ "waist"
175
+ ],
176
+ "mean_std_embedding_keys": null,
177
+ "action_configs": null
178
+ },
179
+ "action": {
180
+ "delta_indices": [
181
+ 0,
182
+ 1,
183
+ 2,
184
+ 3,
185
+ 4,
186
+ 5,
187
+ 6,
188
+ 7,
189
+ 8,
190
+ 9,
191
+ 10,
192
+ 11,
193
+ 12,
194
+ 13,
195
+ 14,
196
+ 15
197
+ ],
198
+ "modality_keys": [
199
+ "left_arm",
200
+ "right_arm",
201
+ "left_hand",
202
+ "right_hand",
203
+ "waist"
204
+ ],
205
+ "sin_cos_embedding_keys": null,
206
+ "mean_std_embedding_keys": null,
207
+ "action_configs": [
208
+ {
209
+ "rep": "RELATIVE",
210
+ "type": "NON_EEF",
211
+ "format": "DEFAULT",
212
+ "state_key": null
213
+ },
214
+ {
215
+ "rep": "RELATIVE",
216
+ "type": "NON_EEF",
217
+ "format": "DEFAULT",
218
+ "state_key": null
219
+ },
220
+ {
221
+ "rep": "RELATIVE",
222
+ "type": "NON_EEF",
223
+ "format": "DEFAULT",
224
+ "state_key": null
225
+ },
226
+ {
227
+ "rep": "RELATIVE",
228
+ "type": "NON_EEF",
229
+ "format": "DEFAULT",
230
+ "state_key": null
231
+ },
232
+ {
233
+ "rep": "ABSOLUTE",
234
+ "type": "NON_EEF",
235
+ "format": "DEFAULT",
236
+ "state_key": null
237
+ }
238
+ ]
239
+ },
240
+ "language": {
241
+ "delta_indices": [
242
+ 0
243
+ ],
244
+ "modality_keys": [
245
+ "task"
246
+ ],
247
+ "sin_cos_embedding_keys": null,
248
+ "mean_std_embedding_keys": null,
249
+ "action_configs": null
250
+ }
251
+ },
252
+ "robocasa_panda_omron": {
253
+ "video": {
254
+ "delta_indices": [
255
+ 0
256
+ ],
257
+ "modality_keys": [
258
+ "res256_image_side_0",
259
+ "res256_image_side_1",
260
+ "res256_image_wrist_0"
261
+ ],
262
+ "sin_cos_embedding_keys": null,
263
+ "mean_std_embedding_keys": null,
264
+ "action_configs": null
265
+ },
266
+ "state": {
267
+ "delta_indices": [
268
+ 0
269
+ ],
270
+ "modality_keys": [
271
+ "end_effector_position_relative",
272
+ "end_effector_rotation_relative",
273
+ "gripper_qpos",
274
+ "base_position",
275
+ "base_rotation"
276
+ ],
277
+ "sin_cos_embedding_keys": null,
278
+ "mean_std_embedding_keys": null,
279
+ "action_configs": null
280
+ },
281
+ "action": {
282
+ "delta_indices": [
283
+ 0,
284
+ 1,
285
+ 2,
286
+ 3,
287
+ 4,
288
+ 5,
289
+ 6,
290
+ 7,
291
+ 8,
292
+ 9,
293
+ 10,
294
+ 11,
295
+ 12,
296
+ 13,
297
+ 14,
298
+ 15
299
+ ],
300
+ "modality_keys": [
301
+ "end_effector_position",
302
+ "end_effector_rotation",
303
+ "gripper_close",
304
+ "base_motion",
305
+ "control_mode"
306
+ ],
307
+ "sin_cos_embedding_keys": null,
308
+ "mean_std_embedding_keys": null,
309
+ "action_configs": [
310
+ {
311
+ "rep": "ABSOLUTE",
312
+ "type": "NON_EEF",
313
+ "format": "DEFAULT",
314
+ "state_key": null
315
+ },
316
+ {
317
+ "rep": "ABSOLUTE",
318
+ "type": "NON_EEF",
319
+ "format": "DEFAULT",
320
+ "state_key": null
321
+ },
322
+ {
323
+ "rep": "ABSOLUTE",
324
+ "type": "NON_EEF",
325
+ "format": "DEFAULT",
326
+ "state_key": null
327
+ },
328
+ {
329
+ "rep": "ABSOLUTE",
330
+ "type": "NON_EEF",
331
+ "format": "DEFAULT",
332
+ "state_key": null
333
+ },
334
+ {
335
+ "rep": "ABSOLUTE",
336
+ "type": "NON_EEF",
337
+ "format": "DEFAULT",
338
+ "state_key": null
339
+ }
340
+ ]
341
+ },
342
+ "language": {
343
+ "delta_indices": [
344
+ 0
345
+ ],
346
+ "modality_keys": [
347
+ "annotation.human.action.task_description"
348
+ ],
349
+ "sin_cos_embedding_keys": null,
350
+ "mean_std_embedding_keys": null,
351
+ "action_configs": null
352
+ }
353
+ },
354
+ "new_embodiment": {
355
+ "video": {
356
+ "delta_indices": [
357
+ 0
358
+ ],
359
+ "modality_keys": [
360
+ "cam_left_head"
361
+ ],
362
+ "sin_cos_embedding_keys": null,
363
+ "mean_std_embedding_keys": null,
364
+ "action_configs": null
365
+ },
366
+ "state": {
367
+ "delta_indices": [
368
+ 0
369
+ ],
370
+ "modality_keys": [
371
+ "arm_left",
372
+ "arm_right"
373
+ ],
374
+ "sin_cos_embedding_keys": null,
375
+ "mean_std_embedding_keys": null,
376
+ "action_configs": null
377
+ },
378
+ "action": {
379
+ "delta_indices": [
380
+ 0,
381
+ 1,
382
+ 2,
383
+ 3,
384
+ 4,
385
+ 5,
386
+ 6,
387
+ 7,
388
+ 8,
389
+ 9,
390
+ 10,
391
+ 11,
392
+ 12,
393
+ 13,
394
+ 14,
395
+ 15
396
+ ],
397
+ "modality_keys": [
398
+ "arm_left",
399
+ "arm_right"
400
+ ],
401
+ "sin_cos_embedding_keys": null,
402
+ "mean_std_embedding_keys": null,
403
+ "action_configs": [
404
+ {
405
+ "rep": "ABSOLUTE",
406
+ "type": "NON_EEF",
407
+ "format": "DEFAULT",
408
+ "state_key": null
409
+ },
410
+ {
411
+ "rep": "ABSOLUTE",
412
+ "type": "NON_EEF",
413
+ "format": "DEFAULT",
414
+ "state_key": null
415
+ }
416
+ ]
417
+ },
418
+ "language": {
419
+ "delta_indices": [
420
+ 0
421
+ ],
422
+ "modality_keys": [
423
+ "annotation.human.task_description"
424
+ ],
425
+ "sin_cos_embedding_keys": null,
426
+ "mean_std_embedding_keys": null,
427
+ "action_configs": null
428
+ }
429
+ }
430
+ },
431
+ "image_crop_size": null,
432
+ "image_target_size": null,
433
+ "use_albumentations": true,
434
+ "random_rotation_angle": null,
435
+ "color_jitter_params": {
436
+ "brightness": 0.3,
437
+ "contrast": 0.4,
438
+ "saturation": 0.5,
439
+ "hue": 0.08
440
+ },
441
+ "shortest_image_edge": 256,
442
+ "crop_fraction": 0.95,
443
+ "model_name": "nvidia/Eagle-Block2A-2B-v2",
444
+ "model_type": "eagle",
445
+ "formalize_language": true,
446
+ "max_state_dim": 128,
447
+ "max_action_dim": 128,
448
+ "max_action_horizon": 50,
449
+ "use_percentiles": false,
450
+ "clip_outliers": true,
451
+ "apply_sincos_state_encoding": true,
452
+ "use_relative_action": true
453
+ }
454
+ }
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:bd3e56ff88f700f7691aef40f8d44ac4dc00d8d2e3e82372e26d0428b6f9961c
3
+ size 5571
wandb_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"project": "finetune-gr00t-n1d6", "run_id": "Box_Packing_0320_lerobot_model"}