Hugging Face's logo

randi233
/
0701-lora-vlm-bs8-50k

Safetensors
Model card Files
xet
 Community
randi233 commited on
Commit
8342e76
·
verified ·
1 Parent(s): 65d18e1

Upload folder using huggingface_hub

Browse files
Files changed (7) hide show
  1. config.json +132 -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 +276 -0
config.json ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "molmoact2",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 6
9
+ ]
10
+ },
11
+ "observation.images.front": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 480,
16
+ 640
17
+ ]
18
+ },
19
+ "observation.images.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
+ 6
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
+ "pretrained_revision": null,
46
+ "checkpoint_path": "allenai/MolmoAct2-SO100_101",
47
+ "checkpoint_revision": null,
48
+ "checkpoint_force_download": false,
49
+ "chunk_size": 10,
50
+ "n_action_steps": 10,
51
+ "action_mode": "continuous",
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 so101 robotic arm",
57
+ "control_mode": "absolute joint pose",
58
+ "image_keys": [
59
+ "observation.images.front",
60
+ "observation.images.wrist"
61
+ ],
62
+ "normalize_language": true,
63
+ "add_setup_tokens": true,
64
+ "add_control_tokens": true,
65
+ "normalize_gripper": true,
66
+ "num_state_tokens": 256,
67
+ "max_sequence_length": null,
68
+ "expected_max_action_dim": 32,
69
+ "num_flow_timesteps": 2,
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": true,
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": 5e-05,
96
+ "optimizer_vit_lr": 5e-05,
97
+ "optimizer_connector_lr": 5e-05,
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": 100000,
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
+ "shoulder_pan.pos",
117
+ "shoulder_lift.pos",
118
+ "elbow_flex.pos",
119
+ "wrist_flex.pos",
120
+ "wrist_roll.pos",
121
+ "gripper.pos"
122
+ ],
123
+ "observation.state": [
124
+ "shoulder_pan.pos",
125
+ "shoulder_lift.pos",
126
+ "elbow_flex.pos",
127
+ "wrist_flex.pos",
128
+ "wrist_roll.pos",
129
+ "gripper.pos"
130
+ ]
131
+ }
132
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d9c2f75c5f0bda248759eccf41157819c1c70172ae08cbb6be04441167bc32f4
3
+ size 11483634784
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
+ 6
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:e2c6bc98fd559abb797c001e69ab5dbedb2a6e7ea9be05e9148738b954201079
3
+ size 7744
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
+ 6
23
+ ]
24
+ },
25
+ "observation.images.front": {
26
+ "type": "VISUAL",
27
+ "shape": [
28
+ 3,
29
+ 480,
30
+ 640
31
+ ]
32
+ },
33
+ "observation.images.wrist": {
34
+ "type": "VISUAL",
35
+ "shape": [
36
+ 3,
37
+ 480,
38
+ 640
39
+ ]
40
+ },
41
+ "action": {
42
+ "type": "ACTION",
43
+ "shape": [
44
+ 6
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-SO100_101",
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.front",
70
+ "observation.images.wrist"
71
+ ],
72
+ "allow_image_key_fallback": false,
73
+ "setup_type": "single so101 robotic arm",
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": 6
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 sha256:555c67edaecac82f8d6a5035531a0aca5173fa96fe1accf6438183845ddb4dd4
3
+ size 7752
train_config.json ADDED
@@ -0,0 +1,276 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "randi233/0609",
4
+ "root": "/workspace/hf-cache/lerobot/randi233/0609",
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": false,
81
+ "streaming": false,
82
+ "eval_split": 0.0
83
+ },
84
+ "env": null,
85
+ "policy": {
86
+ "type": "molmoact2",
87
+ "n_obs_steps": 1,
88
+ "input_features": {
89
+ "observation.state": {
90
+ "type": "STATE",
91
+ "shape": [
92
+ 6
93
+ ]
94
+ },
95
+ "observation.images.front": {
96
+ "type": "VISUAL",
97
+ "shape": [
98
+ 3,
99
+ 480,
100
+ 640
101
+ ]
102
+ },
103
+ "observation.images.wrist": {
104
+ "type": "VISUAL",
105
+ "shape": [
106
+ 3,
107
+ 480,
108
+ 640
109
+ ]
110
+ }
111
+ },
112
+ "output_features": {
113
+ "action": {
114
+ "type": "ACTION",
115
+ "shape": [
116
+ 6
117
+ ]
118
+ }
119
+ },
120
+ "device": "cuda",
121
+ "use_amp": false,
122
+ "use_peft": false,
123
+ "push_to_hub": false,
124
+ "repo_id": null,
125
+ "private": null,
126
+ "tags": null,
127
+ "license": null,
128
+ "pretrained_path": null,
129
+ "pretrained_revision": null,
130
+ "checkpoint_path": "allenai/MolmoAct2-SO100_101",
131
+ "checkpoint_revision": null,
132
+ "checkpoint_force_download": false,
133
+ "chunk_size": 10,
134
+ "n_action_steps": 10,
135
+ "action_mode": "continuous",
136
+ "inference_action_mode": null,
137
+ "discrete_action_tokenizer": "allenai/MolmoAct2-FAST-Tokenizer",
138
+ "discrete_generation_max_steps": null,
139
+ "norm_tag": null,
140
+ "setup_type": "single so101 robotic arm",
141
+ "control_mode": "absolute joint pose",
142
+ "image_keys": [
143
+ "observation.images.front",
144
+ "observation.images.wrist"
145
+ ],
146
+ "normalize_language": true,
147
+ "add_setup_tokens": true,
148
+ "add_control_tokens": true,
149
+ "normalize_gripper": true,
150
+ "num_state_tokens": 256,
151
+ "max_sequence_length": null,
152
+ "expected_max_action_dim": 32,
153
+ "num_flow_timesteps": 2,
154
+ "flow_matching_cutoff": 1.0,
155
+ "flow_matching_time_offset": 0.001,
156
+ "flow_matching_time_scale": 0.999,
157
+ "flow_matching_beta_alpha": 1.0,
158
+ "flow_matching_beta_beta": 1.5,
159
+ "num_inference_steps": null,
160
+ "mask_action_dim_padding": true,
161
+ "enable_inference_cuda_graph": true,
162
+ "per_episode_seed": false,
163
+ "eval_seed": null,
164
+ "rtc_config": null,
165
+ "enable_lora_vlm": true,
166
+ "lora_rank": 64,
167
+ "lora_alpha": 16,
168
+ "lora_dropout": 0.05,
169
+ "lora_bias": "none",
170
+ "enable_lora_action_expert": false,
171
+ "enable_knowledge_insulation": false,
172
+ "freeze_embedding": true,
173
+ "train_action_expert_only": false,
174
+ "gradient_checkpointing": true,
175
+ "model_dtype": "bfloat16",
176
+ "softmax_auxiliary_loss": true,
177
+ "softmax_auxiliary_loss_scale": 0.0001,
178
+ "discrete_loss_token_weighting": "root_subsegments_root_tokens",
179
+ "optimizer_lr": 5e-05,
180
+ "optimizer_vit_lr": 5e-05,
181
+ "optimizer_connector_lr": 5e-05,
182
+ "optimizer_action_expert_lr": 5e-05,
183
+ "optimizer_betas": [
184
+ 0.9,
185
+ 0.95
186
+ ],
187
+ "optimizer_eps": 1e-06,
188
+ "optimizer_weight_decay": 0.0,
189
+ "optimizer_grad_clip_norm": 1.0,
190
+ "scheduler_warmup_steps": 200,
191
+ "scheduler_decay_steps": 100000,
192
+ "scheduler_decay_lr": 1e-06,
193
+ "normalization_mapping": {
194
+ "VISUAL": "IDENTITY",
195
+ "STATE": "QUANTILES",
196
+ "ACTION": "QUANTILES"
197
+ },
198
+ "dataset_feature_names": {
199
+ "action": [
200
+ "shoulder_pan.pos",
201
+ "shoulder_lift.pos",
202
+ "elbow_flex.pos",
203
+ "wrist_flex.pos",
204
+ "wrist_roll.pos",
205
+ "gripper.pos"
206
+ ],
207
+ "observation.state": [
208
+ "shoulder_pan.pos",
209
+ "shoulder_lift.pos",
210
+ "elbow_flex.pos",
211
+ "wrist_flex.pos",
212
+ "wrist_roll.pos",
213
+ "gripper.pos"
214
+ ]
215
+ }
216
+ },
217
+ "reward_model": null,
218
+ "output_dir": "/workspace/outputs/train/0626-lora-vlm-bs8-50k",
219
+ "job_name": "molmoact2_so101_0626_lora_vlm",
220
+ "resume": false,
221
+ "seed": 1000,
222
+ "cudnn_deterministic": false,
223
+ "num_workers": 2,
224
+ "batch_size": 8,
225
+ "prefetch_factor": 4,
226
+ "persistent_workers": true,
227
+ "steps": 50000,
228
+ "env_eval_freq": -1,
229
+ "log_freq": 20,
230
+ "eval_steps": 0,
231
+ "max_eval_samples": 0,
232
+ "tolerance_s": 0.0001,
233
+ "save_checkpoint": true,
234
+ "save_freq": 10000,
235
+ "use_policy_training_preset": true,
236
+ "optimizer": {
237
+ "type": "adamw",
238
+ "lr": 5e-05,
239
+ "weight_decay": 0.0,
240
+ "grad_clip_norm": 1.0,
241
+ "betas": [
242
+ 0.9,
243
+ 0.95
244
+ ],
245
+ "eps": 1e-06
246
+ },
247
+ "scheduler": {
248
+ "type": "cosine_decay_with_warmup",
249
+ "num_warmup_steps": 200,
250
+ "num_decay_steps": 100000,
251
+ "peak_lr": 5e-05,
252
+ "decay_lr": 1e-06
253
+ },
254
+ "eval": {
255
+ "n_episodes": 50,
256
+ "batch_size": 22,
257
+ "use_async_envs": true,
258
+ "recording": false,
259
+ "recording_repo_id": null,
260
+ "recording_private": false
261
+ },
262
+ "wandb": {
263
+ "enable": false,
264
+ "disable_artifact": false,
265
+ "project": "lerobot",
266
+ "entity": null,
267
+ "notes": null,
268
+ "run_id": null,
269
+ "mode": null,
270
+ "add_tags": true
271
+ },
272
+ "peft": null,
273
+ "sample_weighting": null,
274
+ "rename_map": {},
275
+ "checkpoint_path": null
276
+ }