Hugging Face's logo

lirislab
/
scene2_take1

Safetensors
Model card Files
xet
Community
lirislab commited on
Commit
453bd1f
·
verified ·
1 Parent(s): 41499a0

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. config.json +62 -0
  2. model.safetensors +3 -0
  3. train_config.json +233 -0
config.json ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "act",
3
+ "n_obs_steps": 1,
4
+ "normalization_mapping": {
5
+ "VISUAL": "MEAN_STD",
6
+ "STATE": "MEAN_STD",
7
+ "ACTION": "MEAN_STD"
8
+ },
9
+ "input_features": {
10
+ "observation.state": {
11
+ "type": "STATE",
12
+ "shape": [
13
+ 12
14
+ ]
15
+ },
16
+ "observation.images.front": {
17
+ "type": "VISUAL",
18
+ "shape": [
19
+ 3,
20
+ 480,
21
+ 640
22
+ ]
23
+ },
24
+ "observation.images.top": {
25
+ "type": "VISUAL",
26
+ "shape": [
27
+ 3,
28
+ 480,
29
+ 640
30
+ ]
31
+ }
32
+ },
33
+ "output_features": {
34
+ "action": {
35
+ "type": "ACTION",
36
+ "shape": [
37
+ 12
38
+ ]
39
+ }
40
+ },
41
+ "chunk_size": 100,
42
+ "n_action_steps": 100,
43
+ "vision_backbone": "resnet18",
44
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
45
+ "replace_final_stride_with_dilation": false,
46
+ "pre_norm": false,
47
+ "dim_model": 512,
48
+ "n_heads": 8,
49
+ "dim_feedforward": 3200,
50
+ "feedforward_activation": "relu",
51
+ "n_encoder_layers": 4,
52
+ "n_decoder_layers": 1,
53
+ "use_vae": true,
54
+ "latent_dim": 32,
55
+ "n_vae_encoder_layers": 4,
56
+ "temporal_ensemble_coeff": null,
57
+ "dropout": 0.1,
58
+ "kl_weight": 10.0,
59
+ "optimizer_lr": 1e-05,
60
+ "optimizer_weight_decay": 0.0001,
61
+ "optimizer_lr_backbone": 1e-05
62
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b89d5609187770ef0d3c9c6e359e8e6e796baeada9ab9315741fb6a8cd5853d
3
+ size 206750392
train_config.json ADDED
@@ -0,0 +1,233 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "lirislab/bimanual_scene2_take1",
4
+ "episodes": [
5
+ 0,
6
+ 1,
7
+ 2,
8
+ 3,
9
+ 4,
10
+ 5,
11
+ 6,
12
+ 7,
13
+ 8,
14
+ 9,
15
+ 10,
16
+ 12,
17
+ 13,
18
+ 14,
19
+ 15,
20
+ 16,
21
+ 17,
22
+ 18,
23
+ 19,
24
+ 11,
25
+ 20,
26
+ 21,
27
+ 23,
28
+ 24,
29
+ 25,
30
+ 26,
31
+ 27,
32
+ 22,
33
+ 28,
34
+ 29,
35
+ 30,
36
+ 31,
37
+ 32,
38
+ 33,
39
+ 34,
40
+ 35,
41
+ 36,
42
+ 37,
43
+ 38,
44
+ 39,
45
+ 40,
46
+ 41,
47
+ 42,
48
+ 43,
49
+ 44,
50
+ 45,
51
+ 46,
52
+ 47,
53
+ 48,
54
+ 49,
55
+ 50,
56
+ 51
57
+ ],
58
+ "image_transforms": {
59
+ "enable": false,
60
+ "max_num_transforms": 3,
61
+ "random_order": false,
62
+ "tfs": {
63
+ "brightness": {
64
+ "weight": 1.0,
65
+ "type": "ColorJitter",
66
+ "kwargs": {
67
+ "brightness": [
68
+ 0.8,
69
+ 1.2
70
+ ]
71
+ }
72
+ },
73
+ "contrast": {
74
+ "weight": 1.0,
75
+ "type": "ColorJitter",
76
+ "kwargs": {
77
+ "contrast": [
78
+ 0.8,
79
+ 1.2
80
+ ]
81
+ }
82
+ },
83
+ "saturation": {
84
+ "weight": 1.0,
85
+ "type": "ColorJitter",
86
+ "kwargs": {
87
+ "saturation": [
88
+ 0.5,
89
+ 1.5
90
+ ]
91
+ }
92
+ },
93
+ "hue": {
94
+ "weight": 1.0,
95
+ "type": "ColorJitter",
96
+ "kwargs": {
97
+ "hue": [
98
+ -0.05,
99
+ 0.05
100
+ ]
101
+ }
102
+ },
103
+ "sharpness": {
104
+ "weight": 1.0,
105
+ "type": "SharpnessJitter",
106
+ "kwargs": {
107
+ "sharpness": [
108
+ 0.5,
109
+ 1.5
110
+ ]
111
+ }
112
+ }
113
+ }
114
+ },
115
+ "local_files_only": false,
116
+ "use_imagenet_stats": true,
117
+ "video_backend": "pyav"
118
+ },
119
+ "env": null,
120
+ "policy": {
121
+ "type": "act",
122
+ "n_obs_steps": 1,
123
+ "normalization_mapping": {
124
+ "VISUAL": "MEAN_STD",
125
+ "STATE": "MEAN_STD",
126
+ "ACTION": "MEAN_STD"
127
+ },
128
+ "input_features": {
129
+ "observation.state": {
130
+ "type": "STATE",
131
+ "shape": [
132
+ 12
133
+ ]
134
+ },
135
+ "observation.images.front": {
136
+ "type": "VISUAL",
137
+ "shape": [
138
+ 3,
139
+ 480,
140
+ 640
141
+ ]
142
+ },
143
+ "observation.images.top": {
144
+ "type": "VISUAL",
145
+ "shape": [
146
+ 3,
147
+ 480,
148
+ 640
149
+ ]
150
+ }
151
+ },
152
+ "output_features": {
153
+ "action": {
154
+ "type": "ACTION",
155
+ "shape": [
156
+ 12
157
+ ]
158
+ }
159
+ },
160
+ "chunk_size": 100,
161
+ "n_action_steps": 100,
162
+ "vision_backbone": "resnet18",
163
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
164
+ "replace_final_stride_with_dilation": false,
165
+ "pre_norm": false,
166
+ "dim_model": 512,
167
+ "n_heads": 8,
168
+ "dim_feedforward": 3200,
169
+ "feedforward_activation": "relu",
170
+ "n_encoder_layers": 4,
171
+ "n_decoder_layers": 1,
172
+ "use_vae": true,
173
+ "latent_dim": 32,
174
+ "n_vae_encoder_layers": 4,
175
+ "temporal_ensemble_coeff": null,
176
+ "dropout": 0.1,
177
+ "kl_weight": 10.0,
178
+ "optimizer_lr": 1e-05,
179
+ "optimizer_weight_decay": 0.0001,
180
+ "optimizer_lr_backbone": 1e-05
181
+ },
182
+ "output_dir": "outputs/train/act_train_scene2",
183
+ "job_name": "act_so100_bimanual",
184
+ "resume": false,
185
+ "device": "cuda",
186
+ "use_amp": false,
187
+ "seed": 1000,
188
+ "num_workers": 4,
189
+ "batch_size": 8,
190
+ "eval_freq": 20000,
191
+ "log_freq": 200,
192
+ "save_checkpoint": true,
193
+ "save_freq": 20000,
194
+ "offline": {
195
+ "steps": 100000
196
+ },
197
+ "online": {
198
+ "steps": 0,
199
+ "rollout_n_episodes": 1,
200
+ "rollout_batch_size": 1,
201
+ "steps_between_rollouts": null,
202
+ "sampling_ratio": 0.5,
203
+ "env_seed": null,
204
+ "buffer_capacity": null,
205
+ "buffer_seed_size": 0,
206
+ "do_rollout_async": false
207
+ },
208
+ "use_policy_training_preset": true,
209
+ "optimizer": {
210
+ "type": "adamw",
211
+ "lr": 1e-05,
212
+ "weight_decay": 0.0001,
213
+ "grad_clip_norm": 10.0,
214
+ "betas": [
215
+ 0.9,
216
+ 0.999
217
+ ],
218
+ "eps": 1e-08
219
+ },
220
+ "scheduler": null,
221
+ "eval": {
222
+ "n_episodes": 50,
223
+ "batch_size": 50,
224
+ "use_async_envs": false
225
+ },
226
+ "wandb": {
227
+ "enable": true,
228
+ "disable_artifact": false,
229
+ "project": "lerobot",
230
+ "entity": null,
231
+ "notes": null
232
+ }
233
+ }