jlee24 commited on
Commit
f9802d1
·
verified ·
1 Parent(s): c5cf6be

Back up outputs/vla_pickupcube_200_lora/molmoact2_lora_pickupcube_200_20260603_205822/checkpoints/best

Browse files
dk1-teleop/outputs/vla_pickupcube_200_lora/molmoact2_lora_pickupcube_200_20260603_205822/checkpoints/best/checkpoint_manifest.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "checkpoint_format": "trainable_state_dict",
3
+ "step": 446,
4
+ "trainable_tensor_count": 1192,
5
+ "trainable_parameter_count": 614997472,
6
+ "load_note": "Instantiate the base policy/config, then load these tensors with strict=False.",
7
+ "policy_type": "molmoact2"
8
+ }
dk1-teleop/outputs/vla_pickupcube_200_lora/molmoact2_lora_pickupcube_200_20260603_205822/checkpoints/best/pretrained_model/config.json ADDED
@@ -0,0 +1,147 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "molmoact2",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 14
9
+ ]
10
+ },
11
+ "observation.images.top": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 480,
16
+ 640
17
+ ]
18
+ },
19
+ "observation.images.left_wrist": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 480,
24
+ 640
25
+ ]
26
+ }
27
+ },
28
+ "output_features": {
29
+ "action": {
30
+ "type": "ACTION",
31
+ "shape": [
32
+ 14
33
+ ]
34
+ }
35
+ },
36
+ "device": "cuda",
37
+ "use_amp": false,
38
+ "use_peft": false,
39
+ "push_to_hub": false,
40
+ "repo_id": null,
41
+ "private": null,
42
+ "tags": null,
43
+ "license": null,
44
+ "pretrained_path": null,
45
+ "checkpoint_path": "allenai/MolmoAct2",
46
+ "checkpoint_revision": null,
47
+ "checkpoint_force_download": false,
48
+ "chunk_size": 10,
49
+ "n_action_steps": 10,
50
+ "action_mode": "continuous",
51
+ "inference_action_mode": null,
52
+ "discrete_action_tokenizer": "allenai/MolmoAct2-FAST-Tokenizer",
53
+ "discrete_generation_max_steps": null,
54
+ "norm_tag": null,
55
+ "setup_type": "bimanual DK1 robotic arms",
56
+ "control_mode": "absolute joint position control",
57
+ "image_keys": [
58
+ "observation.images.top",
59
+ "observation.images.left_wrist"
60
+ ],
61
+ "normalize_language": true,
62
+ "add_setup_tokens": true,
63
+ "add_control_tokens": true,
64
+ "normalize_gripper": false,
65
+ "num_state_tokens": 256,
66
+ "max_sequence_length": null,
67
+ "expected_max_action_dim": 32,
68
+ "num_flow_timesteps": 8,
69
+ "flow_matching_cutoff": 1.0,
70
+ "flow_matching_time_offset": 0.001,
71
+ "flow_matching_time_scale": 0.999,
72
+ "flow_matching_beta_alpha": 1.0,
73
+ "flow_matching_beta_beta": 1.5,
74
+ "num_inference_steps": null,
75
+ "mask_action_dim_padding": true,
76
+ "enable_inference_cuda_graph": true,
77
+ "per_episode_seed": false,
78
+ "eval_seed": null,
79
+ "rtc_config": null,
80
+ "enable_lora_vlm": true,
81
+ "lora_rank": 16,
82
+ "lora_alpha": 32,
83
+ "lora_dropout": 0.05,
84
+ "lora_bias": "none",
85
+ "enable_lora_action_expert": false,
86
+ "enable_knowledge_insulation": false,
87
+ "freeze_embedding": true,
88
+ "train_action_expert_only": false,
89
+ "gradient_checkpointing": true,
90
+ "model_dtype": "bfloat16",
91
+ "softmax_auxiliary_loss": true,
92
+ "softmax_auxiliary_loss_scale": 0.0001,
93
+ "discrete_loss_token_weighting": "root_subsegments_root_tokens",
94
+ "optimizer_lr": 1e-05,
95
+ "optimizer_vit_lr": 5e-06,
96
+ "optimizer_connector_lr": 5e-06,
97
+ "optimizer_action_expert_lr": 5e-05,
98
+ "optimizer_betas": [
99
+ 0.9,
100
+ 0.95
101
+ ],
102
+ "optimizer_eps": 1e-06,
103
+ "optimizer_weight_decay": 0.0,
104
+ "optimizer_grad_clip_norm": 1.0,
105
+ "scheduler_warmup_steps": 200,
106
+ "scheduler_decay_steps": null,
107
+ "scheduler_decay_lr": 1e-06,
108
+ "normalization_mapping": {
109
+ "VISUAL": "IDENTITY",
110
+ "STATE": "QUANTILES",
111
+ "ACTION": "QUANTILES"
112
+ },
113
+ "dataset_feature_names": {
114
+ "action": [
115
+ "left_joint_1",
116
+ "left_joint_2",
117
+ "left_joint_3",
118
+ "left_joint_4",
119
+ "left_joint_5",
120
+ "left_joint_6",
121
+ "left_gripper",
122
+ "right_joint_1",
123
+ "right_joint_2",
124
+ "right_joint_3",
125
+ "right_joint_4",
126
+ "right_joint_5",
127
+ "right_joint_6",
128
+ "right_gripper"
129
+ ],
130
+ "observation.state": [
131
+ "left_joint_1",
132
+ "left_joint_2",
133
+ "left_joint_3",
134
+ "left_joint_4",
135
+ "left_joint_5",
136
+ "left_joint_6",
137
+ "left_gripper",
138
+ "right_joint_1",
139
+ "right_joint_2",
140
+ "right_joint_3",
141
+ "right_joint_4",
142
+ "right_joint_5",
143
+ "right_joint_6",
144
+ "right_gripper"
145
+ ]
146
+ }
147
+ }
dk1-teleop/outputs/vla_pickupcube_200_lora/molmoact2_lora_pickupcube_200_20260603_205822/checkpoints/best/pretrained_model/policy_postprocessor.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_postprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "molmoact2_clamp_action",
6
+ "config": {}
7
+ },
8
+ {
9
+ "registry_name": "molmoact2_masked_unnormalizer",
10
+ "config": {
11
+ "eps": 1e-08,
12
+ "features": {
13
+ "action": {
14
+ "type": "ACTION",
15
+ "shape": [
16
+ 14
17
+ ]
18
+ }
19
+ },
20
+ "norm_map": {
21
+ "VISUAL": "IDENTITY",
22
+ "STATE": "QUANTILES",
23
+ "ACTION": "QUANTILES"
24
+ }
25
+ },
26
+ "state_file": "policy_postprocessor_step_1_molmoact2_masked_unnormalizer.safetensors"
27
+ },
28
+ {
29
+ "registry_name": "device_processor",
30
+ "config": {
31
+ "device": "cpu",
32
+ "float_dtype": null
33
+ }
34
+ }
35
+ ]
36
+ }
dk1-teleop/outputs/vla_pickupcube_200_lora/molmoact2_lora_pickupcube_200_20260603_205822/checkpoints/best/pretrained_model/policy_postprocessor_step_1_molmoact2_masked_unnormalizer.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:865ec6dd78636d0e63ded6b8e83516a1bbb095c272db56a3d8443ec6ac1ffae1
3
+ size 8496
dk1-teleop/outputs/vla_pickupcube_200_lora/molmoact2_lora_pickupcube_200_20260603_205822/checkpoints/best/pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,93 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_preprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "rename_observations_processor",
6
+ "config": {
7
+ "rename_map": {}
8
+ }
9
+ },
10
+ {
11
+ "registry_name": "to_batch_processor",
12
+ "config": {}
13
+ },
14
+ {
15
+ "registry_name": "molmoact2_masked_normalizer",
16
+ "config": {
17
+ "eps": 1e-08,
18
+ "features": {
19
+ "observation.state": {
20
+ "type": "STATE",
21
+ "shape": [
22
+ 14
23
+ ]
24
+ },
25
+ "observation.images.top": {
26
+ "type": "VISUAL",
27
+ "shape": [
28
+ 3,
29
+ 480,
30
+ 640
31
+ ]
32
+ },
33
+ "observation.images.left_wrist": {
34
+ "type": "VISUAL",
35
+ "shape": [
36
+ 3,
37
+ 480,
38
+ 640
39
+ ]
40
+ },
41
+ "action": {
42
+ "type": "ACTION",
43
+ "shape": [
44
+ 14
45
+ ]
46
+ }
47
+ },
48
+ "norm_map": {
49
+ "VISUAL": "IDENTITY",
50
+ "STATE": "QUANTILES",
51
+ "ACTION": "QUANTILES"
52
+ }
53
+ },
54
+ "state_file": "policy_preprocessor_step_2_molmoact2_masked_normalizer.safetensors"
55
+ },
56
+ {
57
+ "registry_name": "molmoact2_clamp_normalized",
58
+ "config": {}
59
+ },
60
+ {
61
+ "registry_name": "molmoact2_pack_inputs",
62
+ "config": {
63
+ "checkpoint_path": "allenai/MolmoAct2",
64
+ "checkpoint_revision": null,
65
+ "checkpoint_force_download": false,
66
+ "action_mode": "continuous",
67
+ "discrete_action_tokenizer": "allenai/MolmoAct2-FAST-Tokenizer",
68
+ "image_keys": [
69
+ "observation.images.top",
70
+ "observation.images.left_wrist"
71
+ ],
72
+ "allow_image_key_fallback": false,
73
+ "setup_type": "bimanual DK1 robotic arms",
74
+ "control_mode": "absolute joint position control",
75
+ "normalize_language": true,
76
+ "add_setup_tokens": true,
77
+ "add_control_tokens": true,
78
+ "num_state_tokens": 256,
79
+ "max_sequence_length": null,
80
+ "chunk_size": 10,
81
+ "max_action_dim": 32,
82
+ "env_action_dim": 14
83
+ }
84
+ },
85
+ {
86
+ "registry_name": "device_processor",
87
+ "config": {
88
+ "device": "cuda",
89
+ "float_dtype": null
90
+ }
91
+ }
92
+ ]
93
+ }
dk1-teleop/outputs/vla_pickupcube_200_lora/molmoact2_lora_pickupcube_200_20260603_205822/checkpoints/best/pretrained_model/policy_preprocessor_step_2_molmoact2_masked_normalizer.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4103587bbcc5f5881b6dd9a1580250b30b15e1d3dbb2587633c90bb859d2072e
3
+ size 8504
dk1-teleop/outputs/vla_pickupcube_200_lora/molmoact2_lora_pickupcube_200_20260603_205822/checkpoints/best/pretrained_model/train_config.json ADDED
@@ -0,0 +1,285 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "andon/dk1_bimanual_pickupcube_200_10hz_v1_edge_trimmed_2cam_train",
4
+ "root": "/home/andon/.cache/huggingface/lerobot/andon/dk1_bimanual_pickupcube_200_10hz_v1_edge_trimmed_2cam_train",
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": true,
8
+ "max_num_transforms": 3,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "brightness": {
12
+ "weight": 1.0,
13
+ "type": "ColorJitter",
14
+ "kwargs": {
15
+ "brightness": [
16
+ 0.8,
17
+ 1.2
18
+ ]
19
+ }
20
+ },
21
+ "contrast": {
22
+ "weight": 1.0,
23
+ "type": "ColorJitter",
24
+ "kwargs": {
25
+ "contrast": [
26
+ 0.8,
27
+ 1.2
28
+ ]
29
+ }
30
+ },
31
+ "saturation": {
32
+ "weight": 1.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "saturation": [
36
+ 0.5,
37
+ 1.5
38
+ ]
39
+ }
40
+ },
41
+ "hue": {
42
+ "weight": 1.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "hue": [
46
+ -0.05,
47
+ 0.05
48
+ ]
49
+ }
50
+ },
51
+ "sharpness": {
52
+ "weight": 1.0,
53
+ "type": "SharpnessJitter",
54
+ "kwargs": {
55
+ "sharpness": [
56
+ 0.5,
57
+ 1.5
58
+ ]
59
+ }
60
+ },
61
+ "affine": {
62
+ "weight": 1.0,
63
+ "type": "RandomAffine",
64
+ "kwargs": {
65
+ "degrees": [
66
+ -5.0,
67
+ 5.0
68
+ ],
69
+ "translate": [
70
+ 0.05,
71
+ 0.05
72
+ ]
73
+ }
74
+ }
75
+ }
76
+ },
77
+ "revision": null,
78
+ "use_imagenet_stats": true,
79
+ "video_backend": "pyav",
80
+ "return_uint8": true,
81
+ "streaming": false
82
+ },
83
+ "env": null,
84
+ "policy": {
85
+ "type": "molmoact2",
86
+ "n_obs_steps": 1,
87
+ "input_features": {
88
+ "observation.state": {
89
+ "type": "STATE",
90
+ "shape": [
91
+ 14
92
+ ]
93
+ },
94
+ "observation.images.top": {
95
+ "type": "VISUAL",
96
+ "shape": [
97
+ 3,
98
+ 480,
99
+ 640
100
+ ]
101
+ },
102
+ "observation.images.left_wrist": {
103
+ "type": "VISUAL",
104
+ "shape": [
105
+ 3,
106
+ 480,
107
+ 640
108
+ ]
109
+ }
110
+ },
111
+ "output_features": {
112
+ "action": {
113
+ "type": "ACTION",
114
+ "shape": [
115
+ 14
116
+ ]
117
+ }
118
+ },
119
+ "device": "cuda",
120
+ "use_amp": false,
121
+ "use_peft": false,
122
+ "push_to_hub": false,
123
+ "repo_id": null,
124
+ "private": null,
125
+ "tags": null,
126
+ "license": null,
127
+ "pretrained_path": null,
128
+ "checkpoint_path": "allenai/MolmoAct2",
129
+ "checkpoint_revision": null,
130
+ "checkpoint_force_download": false,
131
+ "chunk_size": 10,
132
+ "n_action_steps": 10,
133
+ "action_mode": "continuous",
134
+ "inference_action_mode": null,
135
+ "discrete_action_tokenizer": "allenai/MolmoAct2-FAST-Tokenizer",
136
+ "discrete_generation_max_steps": null,
137
+ "norm_tag": null,
138
+ "setup_type": "bimanual DK1 robotic arms",
139
+ "control_mode": "absolute joint position control",
140
+ "image_keys": [
141
+ "observation.images.top",
142
+ "observation.images.left_wrist"
143
+ ],
144
+ "normalize_language": true,
145
+ "add_setup_tokens": true,
146
+ "add_control_tokens": true,
147
+ "normalize_gripper": false,
148
+ "num_state_tokens": 256,
149
+ "max_sequence_length": null,
150
+ "expected_max_action_dim": 32,
151
+ "num_flow_timesteps": 8,
152
+ "flow_matching_cutoff": 1.0,
153
+ "flow_matching_time_offset": 0.001,
154
+ "flow_matching_time_scale": 0.999,
155
+ "flow_matching_beta_alpha": 1.0,
156
+ "flow_matching_beta_beta": 1.5,
157
+ "num_inference_steps": null,
158
+ "mask_action_dim_padding": true,
159
+ "enable_inference_cuda_graph": true,
160
+ "per_episode_seed": false,
161
+ "eval_seed": null,
162
+ "rtc_config": null,
163
+ "enable_lora_vlm": true,
164
+ "lora_rank": 16,
165
+ "lora_alpha": 32,
166
+ "lora_dropout": 0.05,
167
+ "lora_bias": "none",
168
+ "enable_lora_action_expert": false,
169
+ "enable_knowledge_insulation": false,
170
+ "freeze_embedding": true,
171
+ "train_action_expert_only": false,
172
+ "gradient_checkpointing": true,
173
+ "model_dtype": "bfloat16",
174
+ "softmax_auxiliary_loss": true,
175
+ "softmax_auxiliary_loss_scale": 0.0001,
176
+ "discrete_loss_token_weighting": "root_subsegments_root_tokens",
177
+ "optimizer_lr": 1e-05,
178
+ "optimizer_vit_lr": 5e-06,
179
+ "optimizer_connector_lr": 5e-06,
180
+ "optimizer_action_expert_lr": 5e-05,
181
+ "optimizer_betas": [
182
+ 0.9,
183
+ 0.95
184
+ ],
185
+ "optimizer_eps": 1e-06,
186
+ "optimizer_weight_decay": 0.0,
187
+ "optimizer_grad_clip_norm": 1.0,
188
+ "scheduler_warmup_steps": 200,
189
+ "scheduler_decay_steps": null,
190
+ "scheduler_decay_lr": 1e-06,
191
+ "normalization_mapping": {
192
+ "VISUAL": "IDENTITY",
193
+ "STATE": "QUANTILES",
194
+ "ACTION": "QUANTILES"
195
+ },
196
+ "dataset_feature_names": {
197
+ "action": [
198
+ "left_joint_1",
199
+ "left_joint_2",
200
+ "left_joint_3",
201
+ "left_joint_4",
202
+ "left_joint_5",
203
+ "left_joint_6",
204
+ "left_gripper",
205
+ "right_joint_1",
206
+ "right_joint_2",
207
+ "right_joint_3",
208
+ "right_joint_4",
209
+ "right_joint_5",
210
+ "right_joint_6",
211
+ "right_gripper"
212
+ ],
213
+ "observation.state": [
214
+ "left_joint_1",
215
+ "left_joint_2",
216
+ "left_joint_3",
217
+ "left_joint_4",
218
+ "left_joint_5",
219
+ "left_joint_6",
220
+ "left_gripper",
221
+ "right_joint_1",
222
+ "right_joint_2",
223
+ "right_joint_3",
224
+ "right_joint_4",
225
+ "right_joint_5",
226
+ "right_joint_6",
227
+ "right_gripper"
228
+ ]
229
+ }
230
+ },
231
+ "reward_model": null,
232
+ "output_dir": "/home/andon/clawmachine/dk1-teleop/outputs/vla_pickupcube_200_lora/molmoact2_lora_pickupcube_200_20260603_205822",
233
+ "job_name": "molmoact2_lora_dk1_pickupcube_200_pickupcube_200_20260603_205822",
234
+ "resume": false,
235
+ "seed": 1000,
236
+ "cudnn_deterministic": false,
237
+ "num_workers": 8,
238
+ "batch_size": 16,
239
+ "prefetch_factor": 4,
240
+ "persistent_workers": true,
241
+ "steps": 4460,
242
+ "eval_freq": -1,
243
+ "log_freq": 10,
244
+ "tolerance_s": 0.0001,
245
+ "save_checkpoint": true,
246
+ "save_freq": 0,
247
+ "use_policy_training_preset": false,
248
+ "optimizer": {
249
+ "type": "adamw",
250
+ "lr": 2.5e-05,
251
+ "weight_decay": 0.01,
252
+ "grad_clip_norm": 1.0,
253
+ "betas": [
254
+ 0.9,
255
+ 0.95
256
+ ],
257
+ "eps": 1e-08
258
+ },
259
+ "scheduler": {
260
+ "type": "cosine_decay_with_warmup",
261
+ "num_warmup_steps": 1000,
262
+ "num_decay_steps": 30000,
263
+ "peak_lr": 2.5e-05,
264
+ "decay_lr": 2.5e-06
265
+ },
266
+ "eval": {
267
+ "n_episodes": 50,
268
+ "batch_size": 16,
269
+ "use_async_envs": true
270
+ },
271
+ "wandb": {
272
+ "enable": false,
273
+ "disable_artifact": false,
274
+ "project": "dk1-pickupcube-lora",
275
+ "entity": null,
276
+ "notes": null,
277
+ "run_id": null,
278
+ "mode": null,
279
+ "add_tags": true
280
+ },
281
+ "peft": null,
282
+ "sample_weighting": null,
283
+ "rename_map": {},
284
+ "checkpoint_path": null
285
+ }
dk1-teleop/outputs/vla_pickupcube_200_lora/molmoact2_lora_pickupcube_200_20260603_205822/checkpoints/best/pretrained_model/trainable_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:23dd9be55670554d3b47c9e0d020de817ee607c1e06b2e4963dc58687edd14b7
3
+ size 1305052872
dk1-teleop/outputs/vla_pickupcube_200_lora/molmoact2_lora_pickupcube_200_20260603_205822/checkpoints/best/training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 446
3
+ }