Hugging Face's logo

pepijn223
/
act_aloha_baseline

Safetensors
Model card Files Community
pepijn223 HF Staff commited on
Commit
0a78db0
·
verified ·
1 Parent(s): 08e5a6a

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. config.json +56 -0
  2. model.safetensors +3 -0
  3. train_config.json +195 -0
config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.images.top": {
11
+ "type": "VISUAL",
12
+ "shape": [
13
+ 3,
14
+ 480,
15
+ 640
16
+ ]
17
+ },
18
+ "observation.state": {
19
+ "type": "STATE",
20
+ "shape": [
21
+ 14
22
+ ]
23
+ }
24
+ },
25
+ "output_features": {
26
+ "action": {
27
+ "type": "ACTION",
28
+ "shape": [
29
+ 14
30
+ ]
31
+ }
32
+ },
33
+ "device": "cuda",
34
+ "use_amp": false,
35
+ "chunk_size": 100,
36
+ "n_action_steps": 100,
37
+ "vision_backbone": "resnet18",
38
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
39
+ "replace_final_stride_with_dilation": false,
40
+ "pre_norm": false,
41
+ "dim_model": 512,
42
+ "n_heads": 8,
43
+ "dim_feedforward": 3200,
44
+ "feedforward_activation": "relu",
45
+ "n_encoder_layers": 4,
46
+ "n_decoder_layers": 1,
47
+ "use_vae": true,
48
+ "latent_dim": 32,
49
+ "n_vae_encoder_layers": 4,
50
+ "temporal_ensemble_coeff": null,
51
+ "dropout": 0.1,
52
+ "kl_weight": 10.0,
53
+ "optimizer_lr": 1e-05,
54
+ "optimizer_weight_decay": 0.0001,
55
+ "optimizer_lr_backbone": 1e-05
56
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19df3d3af8dc84596ad662945746230fdd4bc86a879a54999d21a97b7b17c26f
3
+ size 206766560
train_config.json ADDED
@@ -0,0 +1,195 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "lerobot/aloha_sim_transfer_cube_human",
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
+ }
62
+ },
63
+ "revision": null,
64
+ "use_imagenet_stats": true,
65
+ "video_backend": "pyav"
66
+ },
67
+ "env": {
68
+ "type": "aloha",
69
+ "task": "AlohaTransferCube-v0",
70
+ "fps": 50,
71
+ "features": {
72
+ "action": {
73
+ "type": "ACTION",
74
+ "shape": [
75
+ 14
76
+ ]
77
+ },
78
+ "agent_pos": {
79
+ "type": "STATE",
80
+ "shape": [
81
+ 14
82
+ ]
83
+ },
84
+ "pixels/top": {
85
+ "type": "VISUAL",
86
+ "shape": [
87
+ 480,
88
+ 640,
89
+ 3
90
+ ]
91
+ }
92
+ },
93
+ "features_map": {
94
+ "action": "action",
95
+ "agent_pos": "observation.state",
96
+ "top": "observation.image.top",
97
+ "pixels/top": "observation.images.top"
98
+ },
99
+ "episode_length": 400,
100
+ "obs_type": "pixels_agent_pos",
101
+ "render_mode": "rgb_array"
102
+ },
103
+ "policy": {
104
+ "type": "act",
105
+ "n_obs_steps": 1,
106
+ "normalization_mapping": {
107
+ "VISUAL": "MEAN_STD",
108
+ "STATE": "MEAN_STD",
109
+ "ACTION": "MEAN_STD"
110
+ },
111
+ "input_features": {
112
+ "observation.images.top": {
113
+ "type": "VISUAL",
114
+ "shape": [
115
+ 3,
116
+ 480,
117
+ 640
118
+ ]
119
+ },
120
+ "observation.state": {
121
+ "type": "STATE",
122
+ "shape": [
123
+ 14
124
+ ]
125
+ }
126
+ },
127
+ "output_features": {
128
+ "action": {
129
+ "type": "ACTION",
130
+ "shape": [
131
+ 14
132
+ ]
133
+ }
134
+ },
135
+ "device": "cuda",
136
+ "use_amp": false,
137
+ "chunk_size": 100,
138
+ "n_action_steps": 100,
139
+ "vision_backbone": "resnet18",
140
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
141
+ "replace_final_stride_with_dilation": false,
142
+ "pre_norm": false,
143
+ "dim_model": 512,
144
+ "n_heads": 8,
145
+ "dim_feedforward": 3200,
146
+ "feedforward_activation": "relu",
147
+ "n_encoder_layers": 4,
148
+ "n_decoder_layers": 1,
149
+ "use_vae": true,
150
+ "latent_dim": 32,
151
+ "n_vae_encoder_layers": 4,
152
+ "temporal_ensemble_coeff": null,
153
+ "dropout": 0.1,
154
+ "kl_weight": 10.0,
155
+ "optimizer_lr": 1e-05,
156
+ "optimizer_weight_decay": 0.0001,
157
+ "optimizer_lr_backbone": 1e-05
158
+ },
159
+ "output_dir": "outputs/train/act_aloha_transfer",
160
+ "job_name": "aloha_act",
161
+ "resume": false,
162
+ "seed": 1000,
163
+ "num_workers": 4,
164
+ "batch_size": 8,
165
+ "steps": 100000,
166
+ "eval_freq": 20000,
167
+ "log_freq": 200,
168
+ "save_checkpoint": true,
169
+ "save_freq": 20000,
170
+ "use_policy_training_preset": true,
171
+ "optimizer": {
172
+ "type": "adamw",
173
+ "lr": 1e-05,
174
+ "weight_decay": 0.0001,
175
+ "grad_clip_norm": 10.0,
176
+ "betas": [
177
+ 0.9,
178
+ 0.999
179
+ ],
180
+ "eps": 1e-08
181
+ },
182
+ "scheduler": null,
183
+ "eval": {
184
+ "n_episodes": 50,
185
+ "batch_size": 50,
186
+ "use_async_envs": false
187
+ },
188
+ "wandb": {
189
+ "enable": true,
190
+ "disable_artifact": false,
191
+ "project": "lerobot",
192
+ "entity": null,
193
+ "notes": null
194
+ }
195
+ }