Hugging Face's logo

pepijn223
/
sarm_three_500

Safetensors
Model card Files
xet
Community
pepijn223 HF Staff commited on
Commit
1ad5c70
·
verified ·
1 Parent(s): e5dc4b0

Upload folder using huggingface_hub

Browse files
Files changed (6) hide show
  1. config.json +79 -0
  2. model.safetensors +3 -0
  3. policy_postprocessor.json +12 -0
  4. policy_preprocessor.json +61 -0
  5. policy_preprocessor_step_2_normalizer_processor.safetensors +3 -0
  6. train_config.json +213 -0
config.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "sarm",
3
+ "n_obs_steps": 8,
4
+ "input_features": {
5
+ "observation.images.base": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 480,
9
+ 640,
10
+ 3
11
+ ]
12
+ },
13
+ "observation.state": {
14
+ "type": "STATE",
15
+ "shape": [
16
+ 32
17
+ ]
18
+ }
19
+ },
20
+ "output_features": {
21
+ "action": {
22
+ "type": "ACTION",
23
+ "shape": [
24
+ 16
25
+ ]
26
+ }
27
+ },
28
+ "device": "cuda",
29
+ "use_amp": false,
30
+ "push_to_hub": true,
31
+ "repo_id": "pepijn223/sarm_three",
32
+ "private": null,
33
+ "tags": null,
34
+ "license": null,
35
+ "pretrained_path": null,
36
+ "annotation_mode": "dual",
37
+ "frame_gap": 30,
38
+ "max_rewind_steps": 4,
39
+ "image_dim": 512,
40
+ "text_dim": 512,
41
+ "hidden_dim": 768,
42
+ "num_heads": 12,
43
+ "num_layers": 8,
44
+ "max_state_dim": 32,
45
+ "drop_n_last_frames": 1,
46
+ "batch_size": 64,
47
+ "clip_batch_size": 64,
48
+ "dropout": 0.1,
49
+ "stage_loss_weight": 1.0,
50
+ "rewind_probability": 0.8,
51
+ "language_perturbation_probability": 0.2,
52
+ "num_sparse_stages": 1,
53
+ "sparse_subtask_names": [
54
+ "Fold the towel (3 folds in total)"
55
+ ],
56
+ "sparse_temporal_proportions": [
57
+ 1.0
58
+ ],
59
+ "num_dense_stages": 3,
60
+ "dense_subtask_names": [
61
+ "Grab near side and do 1st fold",
62
+ "Grab side and do 2nd fold",
63
+ "Grab side and do 3rd fold to finish folding"
64
+ ],
65
+ "dense_temporal_proportions": [
66
+ 0.24737188507619182,
67
+ 0.29889362942246117,
68
+ 0.45373448550134693
69
+ ],
70
+ "pretrained_model_path": null,
71
+ "image_key": "observation.images.base",
72
+ "state_key": "observation.state",
73
+ "normalization_mapping": {
74
+ "VISUAL": "IDENTITY",
75
+ "STATE": "MEAN_STD",
76
+ "LANGUAGE": "IDENTITY",
77
+ "REWARD": "IDENTITY"
78
+ }
79
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8a6ba5885ed18bff20bb94a989d51ffac0c16fce194dd62ee35cb783240c3402
3
+ size 476745392
policy_postprocessor.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_postprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "device_processor",
6
+ "config": {
7
+ "device": "cpu",
8
+ "float_dtype": null
9
+ }
10
+ }
11
+ ]
12
+ }
policy_preprocessor.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_preprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "to_batch_processor",
6
+ "config": {}
7
+ },
8
+ {
9
+ "registry_name": "rename_observations_processor",
10
+ "config": {
11
+ "rename_map": {}
12
+ }
13
+ },
14
+ {
15
+ "registry_name": "normalizer_processor",
16
+ "config": {
17
+ "eps": 1e-08,
18
+ "features": {
19
+ "observation.images.base": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 480,
23
+ 640,
24
+ 3
25
+ ]
26
+ },
27
+ "observation.state": {
28
+ "type": "STATE",
29
+ "shape": [
30
+ 32
31
+ ]
32
+ },
33
+ "action": {
34
+ "type": "ACTION",
35
+ "shape": [
36
+ 16
37
+ ]
38
+ }
39
+ },
40
+ "norm_map": {
41
+ "VISUAL": "IDENTITY",
42
+ "STATE": "MEAN_STD",
43
+ "LANGUAGE": "IDENTITY",
44
+ "REWARD": "IDENTITY"
45
+ }
46
+ },
47
+ "state_file": "policy_preprocessor_step_2_normalizer_processor.safetensors"
48
+ },
49
+ {
50
+ "class": "lerobot.policies.sarm.processor_sarm.SARMEncodingProcessorStep",
51
+ "config": {}
52
+ },
53
+ {
54
+ "registry_name": "device_processor",
55
+ "config": {
56
+ "device": "cuda",
57
+ "float_dtype": null
58
+ }
59
+ }
60
+ ]
61
+ }
policy_preprocessor_step_2_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5727bdb19fbe881bbc9b55825dc923e0f347ab7d878c33fe36c9a0457bee2d30
3
+ size 9456
train_config.json ADDED
@@ -0,0 +1,213 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "lerobot-data-collection/three-folds-dataset-full-11-07",
4
+ "root": null,
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": false,
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": "torchcodec",
80
+ "streaming": false
81
+ },
82
+ "env": null,
83
+ "policy": {
84
+ "type": "sarm",
85
+ "n_obs_steps": 8,
86
+ "input_features": {
87
+ "observation.images.base": {
88
+ "type": "VISUAL",
89
+ "shape": [
90
+ 480,
91
+ 640,
92
+ 3
93
+ ]
94
+ },
95
+ "observation.state": {
96
+ "type": "STATE",
97
+ "shape": [
98
+ 32
99
+ ]
100
+ }
101
+ },
102
+ "output_features": {
103
+ "action": {
104
+ "type": "ACTION",
105
+ "shape": [
106
+ 16
107
+ ]
108
+ }
109
+ },
110
+ "device": "cuda",
111
+ "use_amp": false,
112
+ "push_to_hub": true,
113
+ "repo_id": "pepijn223/sarm_three",
114
+ "private": null,
115
+ "tags": null,
116
+ "license": null,
117
+ "pretrained_path": null,
118
+ "annotation_mode": "dual",
119
+ "frame_gap": 30,
120
+ "max_rewind_steps": 4,
121
+ "image_dim": 512,
122
+ "text_dim": 512,
123
+ "hidden_dim": 768,
124
+ "num_heads": 12,
125
+ "num_layers": 8,
126
+ "max_state_dim": 32,
127
+ "drop_n_last_frames": 1,
128
+ "batch_size": 64,
129
+ "clip_batch_size": 64,
130
+ "dropout": 0.1,
131
+ "stage_loss_weight": 1.0,
132
+ "rewind_probability": 0.8,
133
+ "language_perturbation_probability": 0.2,
134
+ "num_sparse_stages": 1,
135
+ "sparse_subtask_names": [
136
+ "Fold the towel (3 folds in total)"
137
+ ],
138
+ "sparse_temporal_proportions": [
139
+ 1.0
140
+ ],
141
+ "num_dense_stages": 3,
142
+ "dense_subtask_names": [
143
+ "Grab near side and do 1st fold",
144
+ "Grab side and do 2nd fold",
145
+ "Grab side and do 3rd fold to finish folding"
146
+ ],
147
+ "dense_temporal_proportions": [
148
+ 0.24737188507619182,
149
+ 0.29889362942246117,
150
+ 0.45373448550134693
151
+ ],
152
+ "pretrained_model_path": null,
153
+ "image_key": "observation.images.base",
154
+ "state_key": "observation.state",
155
+ "normalization_mapping": {
156
+ "VISUAL": "IDENTITY",
157
+ "STATE": "MEAN_STD",
158
+ "LANGUAGE": "IDENTITY",
159
+ "REWARD": "IDENTITY"
160
+ }
161
+ },
162
+ "output_dir": "outputs/train/sarm_three",
163
+ "job_name": "sarm_bimanual",
164
+ "resume": false,
165
+ "seed": 1000,
166
+ "num_workers": 4,
167
+ "batch_size": 32,
168
+ "steps": 1000,
169
+ "eval_freq": 20000,
170
+ "log_freq": 200,
171
+ "save_checkpoint": true,
172
+ "save_freq": 500,
173
+ "use_policy_training_preset": true,
174
+ "optimizer": {
175
+ "type": "adamw",
176
+ "lr": 5e-05,
177
+ "weight_decay": 0.001,
178
+ "grad_clip_norm": 10.0,
179
+ "betas": [
180
+ 0.9,
181
+ 0.999
182
+ ],
183
+ "eps": 1e-08
184
+ },
185
+ "scheduler": {
186
+ "type": "cosine_decay_with_warmup",
187
+ "num_warmup_steps": 500,
188
+ "num_decay_steps": 50000,
189
+ "peak_lr": 5e-05,
190
+ "decay_lr": 5e-06
191
+ },
192
+ "eval": {
193
+ "n_episodes": 50,
194
+ "batch_size": 50,
195
+ "use_async_envs": false
196
+ },
197
+ "wandb": {
198
+ "enable": true,
199
+ "disable_artifact": false,
200
+ "project": "sarm",
201
+ "entity": null,
202
+ "notes": null,
203
+ "run_id": "n88a9p7a",
204
+ "mode": null
205
+ },
206
+ "use_rabc": false,
207
+ "rabc_progress_path": null,
208
+ "rabc_kappa": 0.01,
209
+ "rabc_epsilon": 1e-06,
210
+ "rabc_head_mode": "sparse",
211
+ "rename_map": {},
212
+ "checkpoint_path": null
213
+ }