andlyu commited on
Commit
8bb17f2
·
verified ·
1 Parent(s): f5eedc5

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. config.json +64 -0
  2. model.safetensors +3 -0
  3. train_config.json +173 -0
config.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ 6
14
+ ]
15
+ },
16
+ "observation.images.laptop": {
17
+ "type": "VISUAL",
18
+ "shape": [
19
+ 3,
20
+ 480,
21
+ 640
22
+ ]
23
+ },
24
+ "observation.images.phone": {
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
+ 6
38
+ ]
39
+ }
40
+ },
41
+ "device": "cuda",
42
+ "use_amp": false,
43
+ "chunk_size": 100,
44
+ "n_action_steps": 100,
45
+ "vision_backbone": "resnet18",
46
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
47
+ "replace_final_stride_with_dilation": false,
48
+ "pre_norm": false,
49
+ "dim_model": 512,
50
+ "n_heads": 8,
51
+ "dim_feedforward": 3200,
52
+ "feedforward_activation": "relu",
53
+ "n_encoder_layers": 4,
54
+ "n_decoder_layers": 1,
55
+ "use_vae": true,
56
+ "latent_dim": 32,
57
+ "n_vae_encoder_layers": 4,
58
+ "temporal_ensemble_coeff": null,
59
+ "dropout": 0.1,
60
+ "kl_weight": 10.0,
61
+ "optimizer_lr": 1e-05,
62
+ "optimizer_weight_decay": 0.0001,
63
+ "optimizer_lr_backbone": 1e-05
64
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:71dd56b0e53ed202bfeb73cd6a88215f97ba019cc500723433c6a83b657f9bfa
3
+ size 206701072
train_config.json ADDED
@@ -0,0 +1,173 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": [
4
+ "andlyu/lerobot-test-1",
5
+ "andlyu/lerobot-test-1"
6
+ ],
7
+ "root": null,
8
+ "episodes": null,
9
+ "image_transforms": {
10
+ "enable": false,
11
+ "max_num_transforms": 3,
12
+ "random_order": false,
13
+ "tfs": {
14
+ "brightness": {
15
+ "weight": 1.0,
16
+ "type": "ColorJitter",
17
+ "kwargs": {
18
+ "brightness": [
19
+ 0.8,
20
+ 1.2
21
+ ]
22
+ }
23
+ },
24
+ "contrast": {
25
+ "weight": 1.0,
26
+ "type": "ColorJitter",
27
+ "kwargs": {
28
+ "contrast": [
29
+ 0.8,
30
+ 1.2
31
+ ]
32
+ }
33
+ },
34
+ "saturation": {
35
+ "weight": 1.0,
36
+ "type": "ColorJitter",
37
+ "kwargs": {
38
+ "saturation": [
39
+ 0.5,
40
+ 1.5
41
+ ]
42
+ }
43
+ },
44
+ "hue": {
45
+ "weight": 1.0,
46
+ "type": "ColorJitter",
47
+ "kwargs": {
48
+ "hue": [
49
+ -0.05,
50
+ 0.05
51
+ ]
52
+ }
53
+ },
54
+ "sharpness": {
55
+ "weight": 1.0,
56
+ "type": "SharpnessJitter",
57
+ "kwargs": {
58
+ "sharpness": [
59
+ 0.5,
60
+ 1.5
61
+ ]
62
+ }
63
+ }
64
+ }
65
+ },
66
+ "revision": null,
67
+ "use_imagenet_stats": true,
68
+ "video_backend": "torchcodec"
69
+ },
70
+ "env": null,
71
+ "policy": {
72
+ "type": "act",
73
+ "n_obs_steps": 1,
74
+ "normalization_mapping": {
75
+ "VISUAL": "MEAN_STD",
76
+ "STATE": "MEAN_STD",
77
+ "ACTION": "MEAN_STD"
78
+ },
79
+ "input_features": {
80
+ "observation.state": {
81
+ "type": "STATE",
82
+ "shape": [
83
+ 6
84
+ ]
85
+ },
86
+ "observation.images.laptop": {
87
+ "type": "VISUAL",
88
+ "shape": [
89
+ 3,
90
+ 480,
91
+ 640
92
+ ]
93
+ },
94
+ "observation.images.phone": {
95
+ "type": "VISUAL",
96
+ "shape": [
97
+ 3,
98
+ 480,
99
+ 640
100
+ ]
101
+ }
102
+ },
103
+ "output_features": {
104
+ "action": {
105
+ "type": "ACTION",
106
+ "shape": [
107
+ 6
108
+ ]
109
+ }
110
+ },
111
+ "device": "cuda",
112
+ "use_amp": false,
113
+ "chunk_size": 100,
114
+ "n_action_steps": 100,
115
+ "vision_backbone": "resnet18",
116
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
117
+ "replace_final_stride_with_dilation": false,
118
+ "pre_norm": false,
119
+ "dim_model": 512,
120
+ "n_heads": 8,
121
+ "dim_feedforward": 3200,
122
+ "feedforward_activation": "relu",
123
+ "n_encoder_layers": 4,
124
+ "n_decoder_layers": 1,
125
+ "use_vae": true,
126
+ "latent_dim": 32,
127
+ "n_vae_encoder_layers": 4,
128
+ "temporal_ensemble_coeff": null,
129
+ "dropout": 0.1,
130
+ "kl_weight": 10.0,
131
+ "optimizer_lr": 1e-05,
132
+ "optimizer_weight_decay": 0.0001,
133
+ "optimizer_lr_backbone": 1e-05
134
+ },
135
+ "output_dir": "outputs/train/act_so100_canye0",
136
+ "job_name": "act_so100_canye0",
137
+ "resume": false,
138
+ "seed": 1000,
139
+ "num_workers": 4,
140
+ "batch_size": 8,
141
+ "steps": 100000,
142
+ "eval_freq": 20000,
143
+ "log_freq": 200,
144
+ "save_checkpoint": true,
145
+ "save_freq": 10000,
146
+ "use_policy_training_preset": true,
147
+ "optimizer": {
148
+ "type": "adamw",
149
+ "lr": 1e-05,
150
+ "weight_decay": 0.0001,
151
+ "grad_clip_norm": 10.0,
152
+ "betas": [
153
+ 0.9,
154
+ 0.999
155
+ ],
156
+ "eps": 1e-08
157
+ },
158
+ "scheduler": null,
159
+ "eval": {
160
+ "n_episodes": 50,
161
+ "batch_size": 50,
162
+ "use_async_envs": false
163
+ },
164
+ "wandb": {
165
+ "enable": true,
166
+ "disable_artifact": false,
167
+ "project": "lerobot",
168
+ "entity": null,
169
+ "notes": null,
170
+ "run_id": null,
171
+ "mode": null
172
+ }
173
+ }