Hugging Face's logo

jclinton1
/
wedgit_stack_single_vqbet

Safetensors
Model card Files
xet
 Community
2
jclinton1 commited on
Commit
9c0df49
·
verified ·
1 Parent(s): 241f976

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. config.json +73 -0
  2. model.safetensors +3 -0
  3. train_config.json +185 -0
config.json ADDED
@@ -0,0 +1,73 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "vqbet",
3
+ "n_obs_steps": 5,
4
+ "normalization_mapping": {
5
+ "VISUAL": "IDENTITY",
6
+ "STATE": "MIN_MAX",
7
+ "ACTION": "MIN_MAX"
8
+ },
9
+ "input_features": {
10
+ "observation.state": {
11
+ "type": "STATE",
12
+ "shape": [
13
+ 6
14
+ ]
15
+ },
16
+ "observation.images.webcam": {
17
+ "type": "VISUAL",
18
+ "shape": [
19
+ 3,
20
+ 480,
21
+ 640
22
+ ]
23
+ }
24
+ },
25
+ "output_features": {
26
+ "action": {
27
+ "type": "ACTION",
28
+ "shape": [
29
+ 6
30
+ ]
31
+ }
32
+ },
33
+ "device": "cuda",
34
+ "use_amp": false,
35
+ "n_action_pred_token": 3,
36
+ "action_chunk_size": 5,
37
+ "vision_backbone": "resnet18",
38
+ "crop_shape": [
39
+ 84,
40
+ 84
41
+ ],
42
+ "crop_is_random": true,
43
+ "pretrained_backbone_weights": null,
44
+ "use_group_norm": true,
45
+ "spatial_softmax_num_keypoints": 32,
46
+ "n_vqvae_training_steps": 15000,
47
+ "vqvae_n_embed": 16,
48
+ "vqvae_embedding_dim": 256,
49
+ "vqvae_enc_hidden_dim": 128,
50
+ "gpt_block_size": 500,
51
+ "gpt_input_dim": 512,
52
+ "gpt_output_dim": 512,
53
+ "gpt_n_layer": 8,
54
+ "gpt_n_head": 8,
55
+ "gpt_hidden_dim": 512,
56
+ "dropout": 0.1,
57
+ "mlp_hidden_dim": 1024,
58
+ "offset_loss_weight": 10000.0,
59
+ "primary_code_loss_weight": 5.0,
60
+ "secondary_code_loss_weight": 0.5,
61
+ "bet_softmax_temperature": 0.1,
62
+ "sequentially_select": false,
63
+ "optimizer_lr": 0.0001,
64
+ "optimizer_betas": [
65
+ 0.95,
66
+ 0.999
67
+ ],
68
+ "optimizer_eps": 1e-08,
69
+ "optimizer_weight_decay": 1e-06,
70
+ "optimizer_vqvae_lr": 0.001,
71
+ "optimizer_vqvae_weight_decay": 0.0001,
72
+ "scheduler_warmup_steps": 500
73
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c87ebfbab6f00ae3a88976b390149640ac284d4967489ffcb8c6a4ae28704467
3
+ size 159494194
train_config.json ADDED
@@ -0,0 +1,185 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "jclinton1/wedgit_stack_single",
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": null,
68
+ "policy": {
69
+ "type": "vqbet",
70
+ "n_obs_steps": 5,
71
+ "normalization_mapping": {
72
+ "VISUAL": "IDENTITY",
73
+ "STATE": "MIN_MAX",
74
+ "ACTION": "MIN_MAX"
75
+ },
76
+ "input_features": {
77
+ "observation.state": {
78
+ "type": "STATE",
79
+ "shape": [
80
+ 6
81
+ ]
82
+ },
83
+ "observation.images.webcam": {
84
+ "type": "VISUAL",
85
+ "shape": [
86
+ 3,
87
+ 480,
88
+ 640
89
+ ]
90
+ }
91
+ },
92
+ "output_features": {
93
+ "action": {
94
+ "type": "ACTION",
95
+ "shape": [
96
+ 6
97
+ ]
98
+ }
99
+ },
100
+ "device": "cuda",
101
+ "use_amp": false,
102
+ "n_action_pred_token": 3,
103
+ "action_chunk_size": 5,
104
+ "vision_backbone": "resnet18",
105
+ "crop_shape": [
106
+ 84,
107
+ 84
108
+ ],
109
+ "crop_is_random": true,
110
+ "pretrained_backbone_weights": null,
111
+ "use_group_norm": true,
112
+ "spatial_softmax_num_keypoints": 32,
113
+ "n_vqvae_training_steps": 15000,
114
+ "vqvae_n_embed": 16,
115
+ "vqvae_embedding_dim": 256,
116
+ "vqvae_enc_hidden_dim": 128,
117
+ "gpt_block_size": 500,
118
+ "gpt_input_dim": 512,
119
+ "gpt_output_dim": 512,
120
+ "gpt_n_layer": 8,
121
+ "gpt_n_head": 8,
122
+ "gpt_hidden_dim": 512,
123
+ "dropout": 0.1,
124
+ "mlp_hidden_dim": 1024,
125
+ "offset_loss_weight": 10000.0,
126
+ "primary_code_loss_weight": 5.0,
127
+ "secondary_code_loss_weight": 0.5,
128
+ "bet_softmax_temperature": 0.1,
129
+ "sequentially_select": false,
130
+ "optimizer_lr": 0.0001,
131
+ "optimizer_betas": [
132
+ 0.95,
133
+ 0.999
134
+ ],
135
+ "optimizer_eps": 1e-08,
136
+ "optimizer_weight_decay": 1e-06,
137
+ "optimizer_vqvae_lr": 0.001,
138
+ "optimizer_vqvae_weight_decay": 0.0001,
139
+ "scheduler_warmup_steps": 500
140
+ },
141
+ "output_dir": "outputs/train/wedgit_stack_single_val_test2",
142
+ "job_name": "val_test_fixed",
143
+ "resume": false,
144
+ "seed": 1000,
145
+ "num_workers": 8,
146
+ "batch_size": 8,
147
+ "steps": 100000,
148
+ "eval_freq": 20000,
149
+ "validation_freq": 400,
150
+ "log_freq": 200,
151
+ "save_checkpoint": true,
152
+ "save_freq": 20000,
153
+ "use_policy_training_preset": true,
154
+ "optimizer": {
155
+ "type": "adam",
156
+ "lr": 0.0001,
157
+ "weight_decay": 1e-06,
158
+ "grad_clip_norm": 10.0,
159
+ "betas": [
160
+ 0.95,
161
+ 0.999
162
+ ],
163
+ "eps": 1e-08
164
+ },
165
+ "scheduler": {
166
+ "type": "vqbet",
167
+ "num_warmup_steps": 500,
168
+ "num_vqvae_training_steps": 15000,
169
+ "num_cycles": 0.5
170
+ },
171
+ "eval": {
172
+ "n_episodes": 50,
173
+ "batch_size": 50,
174
+ "use_async_envs": false
175
+ },
176
+ "wandb": {
177
+ "enable": true,
178
+ "disable_artifact": false,
179
+ "project": "wedgit_stack_single",
180
+ "entity": null,
181
+ "notes": null,
182
+ "run_id": null,
183
+ "mode": null
184
+ }
185
+ }