Hugging Face's logo

TreeePlanter
/
vls_real_dp_v2

Safetensors
Model card Files
xet
Community
TreeePlanter commited on
Commit
3416deb
·
verified ·
1 Parent(s): 265488e

Upload policy weights and config

Browse files
Files changed (7) hide show
  1. config.json +90 -0
  2. model.safetensors +3 -0
  3. policy_postprocessor.json +32 -0
  4. policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  5. policy_preprocessor.json +64 -0
  6. policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  7. train_config.json +218 -0
config.json ADDED
@@ -0,0 +1,90 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "diffusion",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 8
9
+ ]
10
+ },
11
+ "observation.images.external": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 376,
16
+ 672
17
+ ]
18
+ },
19
+ "observation.images.wrist": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 376,
24
+ 672
25
+ ]
26
+ }
27
+ },
28
+ "output_features": {
29
+ "action": {
30
+ "type": "ACTION",
31
+ "shape": [
32
+ 7
33
+ ]
34
+ }
35
+ },
36
+ "device": "cuda",
37
+ "use_amp": false,
38
+ "push_to_hub": true,
39
+ "repo_id": "lerobot/vls_real_dp_v2",
40
+ "private": null,
41
+ "tags": null,
42
+ "license": null,
43
+ "pretrained_path": null,
44
+ "horizon": 16,
45
+ "n_action_steps": 8,
46
+ "normalization_mapping": {
47
+ "VISUAL": "MEAN_STD",
48
+ "STATE": "MIN_MAX",
49
+ "ACTION": "MIN_MAX"
50
+ },
51
+ "drop_n_last_frames": 7,
52
+ "vision_backbone": "resnet18",
53
+ "crop_shape": [
54
+ 350,
55
+ 640
56
+ ],
57
+ "crop_is_random": true,
58
+ "pretrained_backbone_weights": null,
59
+ "use_group_norm": true,
60
+ "spatial_softmax_num_keypoints": 32,
61
+ "use_separate_rgb_encoder_per_camera": false,
62
+ "down_dims": [
63
+ 256,
64
+ 512,
65
+ 1024
66
+ ],
67
+ "kernel_size": 5,
68
+ "n_groups": 8,
69
+ "diffusion_step_embed_dim": 256,
70
+ "use_film_scale_modulation": true,
71
+ "noise_scheduler_type": "DDPM",
72
+ "num_train_timesteps": 100,
73
+ "beta_schedule": "squaredcos_cap_v2",
74
+ "beta_start": 0.0001,
75
+ "beta_end": 0.02,
76
+ "prediction_type": "epsilon",
77
+ "clip_sample": true,
78
+ "clip_sample_range": 1.0,
79
+ "num_inference_steps": null,
80
+ "do_mask_loss_for_padding": false,
81
+ "optimizer_lr": 0.0001,
82
+ "optimizer_betas": [
83
+ 0.95,
84
+ 0.999
85
+ ],
86
+ "optimizer_eps": 1e-08,
87
+ "optimizer_weight_decay": 1e-06,
88
+ "scheduler_name": "cosine",
89
+ "scheduler_warmup_steps": 500
90
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d2daba8573c571289a021d447ec24ca377a22aafbb32456dc8b4459c402b98
3
+ size 321288092
policy_postprocessor.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_postprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "unnormalizer_processor",
6
+ "config": {
7
+ "eps": 1e-08,
8
+ "features": {
9
+ "action": {
10
+ "type": "ACTION",
11
+ "shape": [
12
+ 7
13
+ ]
14
+ }
15
+ },
16
+ "norm_map": {
17
+ "VISUAL": "MEAN_STD",
18
+ "STATE": "MIN_MAX",
19
+ "ACTION": "MIN_MAX"
20
+ }
21
+ },
22
+ "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors"
23
+ },
24
+ {
25
+ "registry_name": "device_processor",
26
+ "config": {
27
+ "device": "cpu",
28
+ "float_dtype": null
29
+ }
30
+ }
31
+ ]
32
+ }
policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62a98daf0342de2e3b0c48bc2de765f2040f137c52eade0357234089fdcd8643
3
+ size 7692
policy_preprocessor.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_preprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "rename_observations_processor",
6
+ "config": {
7
+ "rename_map": {}
8
+ }
9
+ },
10
+ {
11
+ "registry_name": "to_batch_processor",
12
+ "config": {}
13
+ },
14
+ {
15
+ "registry_name": "device_processor",
16
+ "config": {
17
+ "device": "cuda",
18
+ "float_dtype": null
19
+ }
20
+ },
21
+ {
22
+ "registry_name": "normalizer_processor",
23
+ "config": {
24
+ "eps": 1e-08,
25
+ "features": {
26
+ "observation.state": {
27
+ "type": "STATE",
28
+ "shape": [
29
+ 8
30
+ ]
31
+ },
32
+ "observation.images.external": {
33
+ "type": "VISUAL",
34
+ "shape": [
35
+ 3,
36
+ 376,
37
+ 672
38
+ ]
39
+ },
40
+ "observation.images.wrist": {
41
+ "type": "VISUAL",
42
+ "shape": [
43
+ 3,
44
+ 376,
45
+ 672
46
+ ]
47
+ },
48
+ "action": {
49
+ "type": "ACTION",
50
+ "shape": [
51
+ 7
52
+ ]
53
+ }
54
+ },
55
+ "norm_map": {
56
+ "VISUAL": "MEAN_STD",
57
+ "STATE": "MIN_MAX",
58
+ "ACTION": "MIN_MAX"
59
+ }
60
+ },
61
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
62
+ }
63
+ ]
64
+ }
policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62a98daf0342de2e3b0c48bc2de765f2040f137c52eade0357234089fdcd8643
3
+ size 7692
train_config.json ADDED
@@ -0,0 +1,218 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "TreeePlanter/vls_real_test",
4
+ "repo_ids": null,
5
+ "root": null,
6
+ "episodes": null,
7
+ "image_transforms": {
8
+ "enable": false,
9
+ "max_num_transforms": 3,
10
+ "random_order": false,
11
+ "tfs": {
12
+ "brightness": {
13
+ "weight": 1.0,
14
+ "type": "ColorJitter",
15
+ "kwargs": {
16
+ "brightness": [
17
+ 0.8,
18
+ 1.2
19
+ ]
20
+ }
21
+ },
22
+ "contrast": {
23
+ "weight": 1.0,
24
+ "type": "ColorJitter",
25
+ "kwargs": {
26
+ "contrast": [
27
+ 0.8,
28
+ 1.2
29
+ ]
30
+ }
31
+ },
32
+ "saturation": {
33
+ "weight": 1.0,
34
+ "type": "ColorJitter",
35
+ "kwargs": {
36
+ "saturation": [
37
+ 0.5,
38
+ 1.5
39
+ ]
40
+ }
41
+ },
42
+ "hue": {
43
+ "weight": 1.0,
44
+ "type": "ColorJitter",
45
+ "kwargs": {
46
+ "hue": [
47
+ -0.05,
48
+ 0.05
49
+ ]
50
+ }
51
+ },
52
+ "sharpness": {
53
+ "weight": 1.0,
54
+ "type": "SharpnessJitter",
55
+ "kwargs": {
56
+ "sharpness": [
57
+ 0.5,
58
+ 1.5
59
+ ]
60
+ }
61
+ },
62
+ "affine": {
63
+ "weight": 1.0,
64
+ "type": "RandomAffine",
65
+ "kwargs": {
66
+ "degrees": [
67
+ -5.0,
68
+ 5.0
69
+ ],
70
+ "translate": [
71
+ 0.05,
72
+ 0.05
73
+ ]
74
+ }
75
+ }
76
+ }
77
+ },
78
+ "revision": null,
79
+ "use_imagenet_stats": true,
80
+ "video_backend": "pyav",
81
+ "streaming": false
82
+ },
83
+ "env": null,
84
+ "policy": {
85
+ "type": "diffusion",
86
+ "n_obs_steps": 2,
87
+ "input_features": {
88
+ "observation.state": {
89
+ "type": "STATE",
90
+ "shape": [
91
+ 8
92
+ ]
93
+ },
94
+ "observation.images.external": {
95
+ "type": "VISUAL",
96
+ "shape": [
97
+ 3,
98
+ 376,
99
+ 672
100
+ ]
101
+ },
102
+ "observation.images.wrist": {
103
+ "type": "VISUAL",
104
+ "shape": [
105
+ 3,
106
+ 376,
107
+ 672
108
+ ]
109
+ }
110
+ },
111
+ "output_features": {
112
+ "action": {
113
+ "type": "ACTION",
114
+ "shape": [
115
+ 7
116
+ ]
117
+ }
118
+ },
119
+ "device": "cuda",
120
+ "use_amp": false,
121
+ "push_to_hub": true,
122
+ "repo_id": "lerobot/vls_real_dp_v2",
123
+ "private": null,
124
+ "tags": null,
125
+ "license": null,
126
+ "pretrained_path": null,
127
+ "horizon": 16,
128
+ "n_action_steps": 8,
129
+ "normalization_mapping": {
130
+ "VISUAL": "MEAN_STD",
131
+ "STATE": "MIN_MAX",
132
+ "ACTION": "MIN_MAX"
133
+ },
134
+ "drop_n_last_frames": 7,
135
+ "vision_backbone": "resnet18",
136
+ "crop_shape": [
137
+ 350,
138
+ 640
139
+ ],
140
+ "crop_is_random": true,
141
+ "pretrained_backbone_weights": null,
142
+ "use_group_norm": true,
143
+ "spatial_softmax_num_keypoints": 32,
144
+ "use_separate_rgb_encoder_per_camera": false,
145
+ "down_dims": [
146
+ 256,
147
+ 512,
148
+ 1024
149
+ ],
150
+ "kernel_size": 5,
151
+ "n_groups": 8,
152
+ "diffusion_step_embed_dim": 256,
153
+ "use_film_scale_modulation": true,
154
+ "noise_scheduler_type": "DDPM",
155
+ "num_train_timesteps": 100,
156
+ "beta_schedule": "squaredcos_cap_v2",
157
+ "beta_start": 0.0001,
158
+ "beta_end": 0.02,
159
+ "prediction_type": "epsilon",
160
+ "clip_sample": true,
161
+ "clip_sample_range": 1.0,
162
+ "num_inference_steps": null,
163
+ "do_mask_loss_for_padding": false,
164
+ "optimizer_lr": 0.0001,
165
+ "optimizer_betas": [
166
+ 0.95,
167
+ 0.999
168
+ ],
169
+ "optimizer_eps": 1e-08,
170
+ "optimizer_weight_decay": 1e-06,
171
+ "scheduler_name": "cosine",
172
+ "scheduler_warmup_steps": 500
173
+ },
174
+ "output_dir": "outputs/train/2026-01-21/18-15-44_vls_real_dp_v2",
175
+ "job_name": "vls_real_dp_v2",
176
+ "resume": false,
177
+ "seed": 1000,
178
+ "num_workers": 4,
179
+ "batch_size": 16,
180
+ "steps": 100000,
181
+ "eval_freq": 20000,
182
+ "log_freq": 200,
183
+ "save_checkpoint": true,
184
+ "save_freq": 20000,
185
+ "use_policy_training_preset": true,
186
+ "optimizer": {
187
+ "type": "adam",
188
+ "lr": 0.0001,
189
+ "weight_decay": 1e-06,
190
+ "grad_clip_norm": 10.0,
191
+ "betas": [
192
+ 0.95,
193
+ 0.999
194
+ ],
195
+ "eps": 1e-08
196
+ },
197
+ "scheduler": {
198
+ "type": "diffuser",
199
+ "num_warmup_steps": 500,
200
+ "name": "cosine"
201
+ },
202
+ "eval": {
203
+ "n_episodes": 50,
204
+ "batch_size": 50,
205
+ "use_async_envs": false
206
+ },
207
+ "wandb": {
208
+ "enable": true,
209
+ "disable_artifact": false,
210
+ "project": "lerobot",
211
+ "entity": null,
212
+ "notes": null,
213
+ "run_id": "bdyqv3qd",
214
+ "mode": null
215
+ },
216
+ "checkpoint_path": null,
217
+ "rename_map": {}
218
+ }