Hugging Face's logo

Jiafei1224
/
t3

Safetensors
Model card Files
xet
 Community
Jiafei1224 commited on
Commit
2fc9335
·
verified ·
1 Parent(s): 1bb6e80

Upload checkpoint

Browse files
Files changed (7) hide show
  1. config.json +143 -0
  2. model.safetensors +3 -0
  3. policy_postprocessor.json +36 -0
  4. policy_postprocessor_step_1_molmoact2_masked_unnormalizer.safetensors +3 -0
  5. policy_preprocessor.json +93 -0
  6. policy_preprocessor_step_2_molmoact2_masked_normalizer.safetensors +3 -0
  7. train_config.json +281 -0
config.json ADDED
@@ -0,0 +1,143 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "molmoact2",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 15
9
+ ]
10
+ },
11
+ "observation.images.wrist": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 480,
16
+ 640
17
+ ]
18
+ },
19
+ "observation.images.external": {
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
+ 8
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
+ "trust_remote_code": true,
49
+ "chunk_size": 10,
50
+ "n_action_steps": 10,
51
+ "action_mode": "both",
52
+ "inference_action_mode": null,
53
+ "discrete_action_tokenizer": "allenai/MolmoAct2-FAST-Tokenizer",
54
+ "discrete_generation_max_steps": null,
55
+ "norm_tag": null,
56
+ "setup_type": "single franka robotic arm with FACTR force-feedback teleoperation",
57
+ "control_mode": "absolute joint pose",
58
+ "image_keys": [
59
+ "observation.images.external",
60
+ "observation.images.wrist"
61
+ ],
62
+ "normalize_language": true,
63
+ "add_setup_tokens": true,
64
+ "add_control_tokens": true,
65
+ "normalize_gripper": false,
66
+ "num_state_tokens": 256,
67
+ "max_sequence_length": null,
68
+ "expected_max_action_dim": 32,
69
+ "num_flow_timesteps": 8,
70
+ "flow_matching_cutoff": 1.0,
71
+ "flow_matching_time_offset": 0.001,
72
+ "flow_matching_time_scale": 0.999,
73
+ "flow_matching_beta_alpha": 1.0,
74
+ "flow_matching_beta_beta": 1.5,
75
+ "num_inference_steps": null,
76
+ "mask_action_dim_padding": true,
77
+ "enable_inference_cuda_graph": true,
78
+ "per_episode_seed": false,
79
+ "eval_seed": null,
80
+ "rtc_config": null,
81
+ "enable_lora_vlm": false,
82
+ "lora_rank": 64,
83
+ "lora_alpha": 16,
84
+ "lora_dropout": 0.05,
85
+ "lora_bias": "none",
86
+ "enable_lora_action_expert": false,
87
+ "enable_knowledge_insulation": false,
88
+ "freeze_embedding": true,
89
+ "train_action_expert_only": false,
90
+ "gradient_checkpointing": true,
91
+ "model_dtype": "bfloat16",
92
+ "softmax_auxiliary_loss": true,
93
+ "softmax_auxiliary_loss_scale": 0.0001,
94
+ "discrete_loss_token_weighting": "root_subsegments_root_tokens",
95
+ "optimizer_lr": 1e-05,
96
+ "optimizer_vit_lr": 5e-06,
97
+ "optimizer_connector_lr": 5e-06,
98
+ "optimizer_action_expert_lr": 5e-05,
99
+ "optimizer_betas": [
100
+ 0.9,
101
+ 0.95
102
+ ],
103
+ "optimizer_eps": 1e-06,
104
+ "optimizer_weight_decay": 0.0,
105
+ "optimizer_grad_clip_norm": 1.0,
106
+ "scheduler_warmup_steps": 200,
107
+ "scheduler_decay_steps": null,
108
+ "scheduler_decay_lr": 1e-06,
109
+ "normalization_mapping": {
110
+ "VISUAL": "IDENTITY",
111
+ "STATE": "QUANTILES",
112
+ "ACTION": "QUANTILES"
113
+ },
114
+ "dataset_feature_names": {
115
+ "action": [
116
+ "q_cmd_0",
117
+ "q_cmd_1",
118
+ "q_cmd_2",
119
+ "q_cmd_3",
120
+ "q_cmd_4",
121
+ "q_cmd_5",
122
+ "q_cmd_6",
123
+ "gripper_cmd"
124
+ ],
125
+ "observation.state": [
126
+ "q_0",
127
+ "q_1",
128
+ "q_2",
129
+ "q_3",
130
+ "q_4",
131
+ "q_5",
132
+ "q_6",
133
+ "dq_0",
134
+ "dq_1",
135
+ "dq_2",
136
+ "dq_3",
137
+ "dq_4",
138
+ "dq_5",
139
+ "dq_6",
140
+ "gripper_width"
141
+ ]
142
+ }
143
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff93f48a9da1d2bb863fb9fc4e6c1984608c558aadf0f60b608bf15c7b138360
3
+ size 10884573720
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
+ 8
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
+ }
policy_postprocessor_step_1_molmoact2_masked_unnormalizer.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5fae1aacc548354e41185579ab4c043103cdbbb25c833cd9243cf4c205031c9f
3
+ size 8280
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
+ 15
23
+ ]
24
+ },
25
+ "observation.images.wrist": {
26
+ "type": "VISUAL",
27
+ "shape": [
28
+ 3,
29
+ 480,
30
+ 640
31
+ ]
32
+ },
33
+ "observation.images.external": {
34
+ "type": "VISUAL",
35
+ "shape": [
36
+ 3,
37
+ 480,
38
+ 640
39
+ ]
40
+ },
41
+ "action": {
42
+ "type": "ACTION",
43
+ "shape": [
44
+ 8
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
+ "trust_remote_code": true,
67
+ "action_mode": "both",
68
+ "discrete_action_tokenizer": "allenai/MolmoAct2-FAST-Tokenizer",
69
+ "image_keys": [
70
+ "observation.images.external",
71
+ "observation.images.wrist"
72
+ ],
73
+ "setup_type": "single franka robotic arm with FACTR force-feedback teleoperation",
74
+ "control_mode": "absolute joint pose",
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": 8
83
+ }
84
+ },
85
+ {
86
+ "registry_name": "device_processor",
87
+ "config": {
88
+ "device": "cuda",
89
+ "float_dtype": null
90
+ }
91
+ }
92
+ ]
93
+ }
policy_preprocessor_step_2_molmoact2_masked_normalizer.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5fae1aacc548354e41185579ab4c043103cdbbb25c833cd9243cf4c205031c9f
3
+ size 8280
train_config.json ADDED
@@ -0,0 +1,281 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "Yuxin-Li/franka_factr_teleop",
4
+ "root": null,
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": "main",
78
+ "use_imagenet_stats": true,
79
+ "video_backend": "pyav",
80
+ "return_uint8": false,
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
+ 15
92
+ ]
93
+ },
94
+ "observation.images.wrist": {
95
+ "type": "VISUAL",
96
+ "shape": [
97
+ 3,
98
+ 480,
99
+ 640
100
+ ]
101
+ },
102
+ "observation.images.external": {
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
+ 8
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
+ "trust_remote_code": true,
132
+ "chunk_size": 10,
133
+ "n_action_steps": 10,
134
+ "action_mode": "both",
135
+ "inference_action_mode": null,
136
+ "discrete_action_tokenizer": "allenai/MolmoAct2-FAST-Tokenizer",
137
+ "discrete_generation_max_steps": null,
138
+ "norm_tag": null,
139
+ "setup_type": "single franka robotic arm with FACTR force-feedback teleoperation",
140
+ "control_mode": "absolute joint pose",
141
+ "image_keys": [
142
+ "observation.images.external",
143
+ "observation.images.wrist"
144
+ ],
145
+ "normalize_language": true,
146
+ "add_setup_tokens": true,
147
+ "add_control_tokens": true,
148
+ "normalize_gripper": false,
149
+ "num_state_tokens": 256,
150
+ "max_sequence_length": null,
151
+ "expected_max_action_dim": 32,
152
+ "num_flow_timesteps": 8,
153
+ "flow_matching_cutoff": 1.0,
154
+ "flow_matching_time_offset": 0.001,
155
+ "flow_matching_time_scale": 0.999,
156
+ "flow_matching_beta_alpha": 1.0,
157
+ "flow_matching_beta_beta": 1.5,
158
+ "num_inference_steps": null,
159
+ "mask_action_dim_padding": true,
160
+ "enable_inference_cuda_graph": true,
161
+ "per_episode_seed": false,
162
+ "eval_seed": null,
163
+ "rtc_config": null,
164
+ "enable_lora_vlm": false,
165
+ "lora_rank": 64,
166
+ "lora_alpha": 16,
167
+ "lora_dropout": 0.05,
168
+ "lora_bias": "none",
169
+ "enable_lora_action_expert": false,
170
+ "enable_knowledge_insulation": false,
171
+ "freeze_embedding": true,
172
+ "train_action_expert_only": false,
173
+ "gradient_checkpointing": true,
174
+ "model_dtype": "bfloat16",
175
+ "softmax_auxiliary_loss": true,
176
+ "softmax_auxiliary_loss_scale": 0.0001,
177
+ "discrete_loss_token_weighting": "root_subsegments_root_tokens",
178
+ "optimizer_lr": 1e-05,
179
+ "optimizer_vit_lr": 5e-06,
180
+ "optimizer_connector_lr": 5e-06,
181
+ "optimizer_action_expert_lr": 5e-05,
182
+ "optimizer_betas": [
183
+ 0.9,
184
+ 0.95
185
+ ],
186
+ "optimizer_eps": 1e-06,
187
+ "optimizer_weight_decay": 0.0,
188
+ "optimizer_grad_clip_norm": 1.0,
189
+ "scheduler_warmup_steps": 200,
190
+ "scheduler_decay_steps": null,
191
+ "scheduler_decay_lr": 1e-06,
192
+ "normalization_mapping": {
193
+ "VISUAL": "IDENTITY",
194
+ "STATE": "QUANTILES",
195
+ "ACTION": "QUANTILES"
196
+ },
197
+ "dataset_feature_names": {
198
+ "action": [
199
+ "q_cmd_0",
200
+ "q_cmd_1",
201
+ "q_cmd_2",
202
+ "q_cmd_3",
203
+ "q_cmd_4",
204
+ "q_cmd_5",
205
+ "q_cmd_6",
206
+ "gripper_cmd"
207
+ ],
208
+ "observation.state": [
209
+ "q_0",
210
+ "q_1",
211
+ "q_2",
212
+ "q_3",
213
+ "q_4",
214
+ "q_5",
215
+ "q_6",
216
+ "dq_0",
217
+ "dq_1",
218
+ "dq_2",
219
+ "dq_3",
220
+ "dq_4",
221
+ "dq_5",
222
+ "dq_6",
223
+ "gripper_width"
224
+ ]
225
+ }
226
+ },
227
+ "reward_model": null,
228
+ "output_dir": "outputs/molmoact2_franka_factr_ft",
229
+ "job_name": "molmoact2_franka_factr_ft",
230
+ "resume": false,
231
+ "seed": 1000,
232
+ "cudnn_deterministic": false,
233
+ "num_workers": 4,
234
+ "batch_size": 4,
235
+ "prefetch_factor": 4,
236
+ "persistent_workers": true,
237
+ "steps": 10000,
238
+ "eval_freq": -1,
239
+ "log_freq": 20,
240
+ "tolerance_s": 0.0001,
241
+ "save_checkpoint": true,
242
+ "save_freq": 2000,
243
+ "use_policy_training_preset": true,
244
+ "optimizer": {
245
+ "type": "adamw",
246
+ "lr": 1e-05,
247
+ "weight_decay": 0.0,
248
+ "grad_clip_norm": 1.0,
249
+ "betas": [
250
+ 0.9,
251
+ 0.95
252
+ ],
253
+ "eps": 1e-06
254
+ },
255
+ "scheduler": {
256
+ "type": "molmoact2_cosine_decay_with_warmup",
257
+ "num_warmup_steps": 200,
258
+ "num_decay_steps": null,
259
+ "peak_lr": 1e-05,
260
+ "decay_lr": 1e-06
261
+ },
262
+ "eval": {
263
+ "n_episodes": 50,
264
+ "batch_size": 50,
265
+ "use_async_envs": true
266
+ },
267
+ "wandb": {
268
+ "enable": false,
269
+ "disable_artifact": false,
270
+ "project": "molmoact2-finetune",
271
+ "entity": "my-lab",
272
+ "notes": null,
273
+ "run_id": null,
274
+ "mode": null,
275
+ "add_tags": true
276
+ },
277
+ "peft": null,
278
+ "sample_weighting": null,
279
+ "rename_map": {},
280
+ "checkpoint_path": null
281
+ }