Hugging Face's logo

ssaito
/
act_pusht_test

Safetensors
Model card Files
xet
 Community
ssaito commited on
Commit
ea1366d
·
verified ·
1 Parent(s): 0923255

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 +206 -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.image": {
11
+ "type": "VISUAL",
12
+ "shape": [
13
+ 3,
14
+ 96,
15
+ 96
16
+ ]
17
+ },
18
+ "observation.state": {
19
+ "type": "STATE",
20
+ "shape": [
21
+ 2
22
+ ]
23
+ }
24
+ },
25
+ "output_features": {
26
+ "action": {
27
+ "type": "ACTION",
28
+ "shape": [
29
+ 2
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 sha256:2a4c50e519e7ed60f787226d78b8f0b44f8f0bba05d5ec45e5919cb85420d56f
3
+ size 206667888
train_config.json ADDED
@@ -0,0 +1,206 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "lerobot/pusht",
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": "torchcodec"
66
+ },
67
+ "env": {
68
+ "type": "pusht",
69
+ "task": "PushT-v0",
70
+ "fps": 10,
71
+ "features": {
72
+ "action": {
73
+ "type": "ACTION",
74
+ "shape": [
75
+ 2
76
+ ]
77
+ },
78
+ "agent_pos": {
79
+ "type": "STATE",
80
+ "shape": [
81
+ 2
82
+ ]
83
+ },
84
+ "pixels": {
85
+ "type": "VISUAL",
86
+ "shape": [
87
+ 384,
88
+ 384,
89
+ 3
90
+ ]
91
+ }
92
+ },
93
+ "features_map": {
94
+ "action": "action",
95
+ "agent_pos": "observation.state",
96
+ "environment_state": "observation.environment_state",
97
+ "pixels": "observation.image"
98
+ },
99
+ "episode_length": 300,
100
+ "obs_type": "pixels_agent_pos",
101
+ "render_mode": "rgb_array",
102
+ "visualization_width": 384,
103
+ "visualization_height": 384
104
+ },
105
+ "policy": {
106
+ "type": "act",
107
+ "n_obs_steps": 1,
108
+ "normalization_mapping": {
109
+ "VISUAL": "MEAN_STD",
110
+ "STATE": "MEAN_STD",
111
+ "ACTION": "MEAN_STD"
112
+ },
113
+ "input_features": {
114
+ "observation.image": {
115
+ "type": "VISUAL",
116
+ "shape": [
117
+ 3,
118
+ 96,
119
+ 96
120
+ ]
121
+ },
122
+ "observation.state": {
123
+ "type": "STATE",
124
+ "shape": [
125
+ 2
126
+ ]
127
+ }
128
+ },
129
+ "output_features": {
130
+ "action": {
131
+ "type": "ACTION",
132
+ "shape": [
133
+ 2
134
+ ]
135
+ }
136
+ },
137
+ "device": "cuda",
138
+ "use_amp": false,
139
+ "chunk_size": 100,
140
+ "n_action_steps": 100,
141
+ "vision_backbone": "resnet18",
142
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
143
+ "replace_final_stride_with_dilation": false,
144
+ "pre_norm": false,
145
+ "dim_model": 512,
146
+ "n_heads": 8,
147
+ "dim_feedforward": 3200,
148
+ "feedforward_activation": "relu",
149
+ "n_encoder_layers": 4,
150
+ "n_decoder_layers": 1,
151
+ "use_vae": true,
152
+ "latent_dim": 32,
153
+ "n_vae_encoder_layers": 4,
154
+ "temporal_ensemble_coeff": null,
155
+ "dropout": 0.1,
156
+ "kl_weight": 10.0,
157
+ "optimizer_lr": 1e-05,
158
+ "optimizer_weight_decay": 0.0001,
159
+ "optimizer_lr_backbone": 1e-05
160
+ },
161
+ "output_dir": "outputs/train/act_pusht_test",
162
+ "job_name": "pusht_act",
163
+ "resume": false,
164
+ "seed": 1000,
165
+ "num_workers": 4,
166
+ "batch_size": 8,
167
+ "steps": 100000,
168
+ "eval_freq": 20000,
169
+ "log_freq": 200,
170
+ "save_checkpoint": true,
171
+ "save_freq": 20000,
172
+ "use_policy_training_preset": true,
173
+ "optimizer": {
174
+ "type": "adamw",
175
+ "lr": 1e-05,
176
+ "weight_decay": 0.0001,
177
+ "grad_clip_norm": 10.0,
178
+ "betas": [
179
+ 0.9,
180
+ 0.999
181
+ ],
182
+ "eps": 1e-08
183
+ },
184
+ "scheduler": null,
185
+ "eval": {
186
+ "n_episodes": 50,
187
+ "batch_size": 50,
188
+ "use_async_envs": false
189
+ },
190
+ "wandb": {
191
+ "enable": false,
192
+ "disable_artifact": false,
193
+ "project": "lerobot",
194
+ "entity": null,
195
+ "notes": null,
196
+ "run_id": null,
197
+ "mode": null
198
+ },
199
+ "tensorboard": {
200
+ "enable": true,
201
+ "log_dir": null,
202
+ "comment": null,
203
+ "flush_secs": 120,
204
+ "disable_artifact": false
205
+ }
206
+ }