Hugging Face's logo

ywu67
/
b1k-task0000

Safetensors
Model card Files
xet
Community
ywu67 commited on
Commit
30a9b5c
·
verified ·
1 Parent(s): d80ce93

Upload checkpoint at step 15000

Browse files
Files changed (7) hide show
  1. checkpoint-15000/config.json +138 -0
  2. checkpoint-15000/model.safetensors +3 -0
  3. checkpoint-15000/policy_postprocessor.json +32 -0
  4. checkpoint-15000/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  5. checkpoint-15000/policy_preprocessor.json +132 -0
  6. checkpoint-15000/policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  7. checkpoint-15000/train_config.json +262 -0
checkpoint-15000/config.json ADDED
@@ -0,0 +1,138 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "act",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.images.rgb.left_wrist": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 480,
9
+ 480,
10
+ 3
11
+ ]
12
+ },
13
+ "observation.images.rgb.right_wrist": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 480,
17
+ 480,
18
+ 3
19
+ ]
20
+ },
21
+ "observation.images.rgb.head": {
22
+ "type": "VISUAL",
23
+ "shape": [
24
+ 720,
25
+ 720,
26
+ 3
27
+ ]
28
+ },
29
+ "observation.images.depth.left_wrist": {
30
+ "type": "VISUAL",
31
+ "shape": [
32
+ 480,
33
+ 480,
34
+ 3
35
+ ]
36
+ },
37
+ "observation.images.depth.right_wrist": {
38
+ "type": "VISUAL",
39
+ "shape": [
40
+ 480,
41
+ 480,
42
+ 3
43
+ ]
44
+ },
45
+ "observation.images.depth.head": {
46
+ "type": "VISUAL",
47
+ "shape": [
48
+ 720,
49
+ 720,
50
+ 3
51
+ ]
52
+ },
53
+ "observation.images.seg_instance_id.left_wrist": {
54
+ "type": "VISUAL",
55
+ "shape": [
56
+ 480,
57
+ 480,
58
+ 3
59
+ ]
60
+ },
61
+ "observation.images.seg_instance_id.right_wrist": {
62
+ "type": "VISUAL",
63
+ "shape": [
64
+ 480,
65
+ 480,
66
+ 3
67
+ ]
68
+ },
69
+ "observation.images.seg_instance_id.head": {
70
+ "type": "VISUAL",
71
+ "shape": [
72
+ 720,
73
+ 720,
74
+ 3
75
+ ]
76
+ },
77
+ "observation.cam_rel_poses": {
78
+ "type": "STATE",
79
+ "shape": [
80
+ 21
81
+ ]
82
+ },
83
+ "observation.state": {
84
+ "type": "STATE",
85
+ "shape": [
86
+ 256
87
+ ]
88
+ },
89
+ "observation.task_info": {
90
+ "type": "STATE",
91
+ "shape": [
92
+ 46
93
+ ]
94
+ }
95
+ },
96
+ "output_features": {
97
+ "action": {
98
+ "type": "ACTION",
99
+ "shape": [
100
+ 23
101
+ ]
102
+ }
103
+ },
104
+ "device": "cuda",
105
+ "use_amp": true,
106
+ "push_to_hub": true,
107
+ "repo_id": "ywu67/b1k-task0000",
108
+ "private": null,
109
+ "tags": null,
110
+ "license": null,
111
+ "pretrained_path": null,
112
+ "chunk_size": 100,
113
+ "n_action_steps": 100,
114
+ "normalization_mapping": {
115
+ "VISUAL": "MEAN_STD",
116
+ "STATE": "MEAN_STD",
117
+ "ACTION": "MEAN_STD"
118
+ },
119
+ "vision_backbone": "resnet18",
120
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
121
+ "replace_final_stride_with_dilation": false,
122
+ "pre_norm": false,
123
+ "dim_model": 512,
124
+ "n_heads": 8,
125
+ "dim_feedforward": 3200,
126
+ "feedforward_activation": "relu",
127
+ "n_encoder_layers": 4,
128
+ "n_decoder_layers": 1,
129
+ "use_vae": true,
130
+ "latent_dim": 32,
131
+ "n_vae_encoder_layers": 4,
132
+ "temporal_ensemble_coeff": null,
133
+ "dropout": 0.1,
134
+ "kl_weight": 10.0,
135
+ "optimizer_lr": 1e-05,
136
+ "optimizer_weight_decay": 0.0001,
137
+ "optimizer_lr_backbone": 1e-05
138
+ }
checkpoint-15000/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e0349d808bb13128976544712c7b00ec85baa8656b3214c36b6471006e9d6f28
3
+ size 207793444
checkpoint-15000/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
+ 23
13
+ ]
14
+ }
15
+ },
16
+ "norm_map": {
17
+ "VISUAL": "MEAN_STD",
18
+ "STATE": "MEAN_STD",
19
+ "ACTION": "MEAN_STD"
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
+ }
checkpoint-15000/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:089d5ed6f26ba87471187b591e4b793af3c976b96dabb5912b5ee36591a2770f
3
+ size 10924
checkpoint-15000/policy_preprocessor.json ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.images.rgb.left_wrist": {
27
+ "type": "VISUAL",
28
+ "shape": [
29
+ 480,
30
+ 480,
31
+ 3
32
+ ]
33
+ },
34
+ "observation.images.rgb.right_wrist": {
35
+ "type": "VISUAL",
36
+ "shape": [
37
+ 480,
38
+ 480,
39
+ 3
40
+ ]
41
+ },
42
+ "observation.images.rgb.head": {
43
+ "type": "VISUAL",
44
+ "shape": [
45
+ 720,
46
+ 720,
47
+ 3
48
+ ]
49
+ },
50
+ "observation.images.depth.left_wrist": {
51
+ "type": "VISUAL",
52
+ "shape": [
53
+ 480,
54
+ 480,
55
+ 3
56
+ ]
57
+ },
58
+ "observation.images.depth.right_wrist": {
59
+ "type": "VISUAL",
60
+ "shape": [
61
+ 480,
62
+ 480,
63
+ 3
64
+ ]
65
+ },
66
+ "observation.images.depth.head": {
67
+ "type": "VISUAL",
68
+ "shape": [
69
+ 720,
70
+ 720,
71
+ 3
72
+ ]
73
+ },
74
+ "observation.images.seg_instance_id.left_wrist": {
75
+ "type": "VISUAL",
76
+ "shape": [
77
+ 480,
78
+ 480,
79
+ 3
80
+ ]
81
+ },
82
+ "observation.images.seg_instance_id.right_wrist": {
83
+ "type": "VISUAL",
84
+ "shape": [
85
+ 480,
86
+ 480,
87
+ 3
88
+ ]
89
+ },
90
+ "observation.images.seg_instance_id.head": {
91
+ "type": "VISUAL",
92
+ "shape": [
93
+ 720,
94
+ 720,
95
+ 3
96
+ ]
97
+ },
98
+ "observation.cam_rel_poses": {
99
+ "type": "STATE",
100
+ "shape": [
101
+ 21
102
+ ]
103
+ },
104
+ "observation.state": {
105
+ "type": "STATE",
106
+ "shape": [
107
+ 256
108
+ ]
109
+ },
110
+ "observation.task_info": {
111
+ "type": "STATE",
112
+ "shape": [
113
+ 46
114
+ ]
115
+ },
116
+ "action": {
117
+ "type": "ACTION",
118
+ "shape": [
119
+ 23
120
+ ]
121
+ }
122
+ },
123
+ "norm_map": {
124
+ "VISUAL": "MEAN_STD",
125
+ "STATE": "MEAN_STD",
126
+ "ACTION": "MEAN_STD"
127
+ }
128
+ },
129
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
130
+ }
131
+ ]
132
+ }
checkpoint-15000/policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:089d5ed6f26ba87471187b591e4b793af3c976b96dabb5912b5ee36591a2770f
3
+ size 10924
checkpoint-15000/train_config.json ADDED
@@ -0,0 +1,262 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "ywu67/2025-challenge-demos-task0000",
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
+ "affine": {
62
+ "weight": 1.0,
63
+ "type": "RandomAffine",
64
+ "kwargs": {
65
+ "degrees": [
66
+ -5.0,
67
+ 5.0
68
+ ],
69
+ "translate": [
70
+ 0.05,
71
+ 0.05
72
+ ]
73
+ }
74
+ }
75
+ }
76
+ },
77
+ "revision": null,
78
+ "use_imagenet_stats": true,
79
+ "video_backend": "torchcodec",
80
+ "streaming": false,
81
+ "tolerance_s": 0.0001
82
+ },
83
+ "env": null,
84
+ "policy": {
85
+ "type": "act",
86
+ "n_obs_steps": 1,
87
+ "input_features": {
88
+ "observation.images.rgb.left_wrist": {
89
+ "type": "VISUAL",
90
+ "shape": [
91
+ 480,
92
+ 480,
93
+ 3
94
+ ]
95
+ },
96
+ "observation.images.rgb.right_wrist": {
97
+ "type": "VISUAL",
98
+ "shape": [
99
+ 480,
100
+ 480,
101
+ 3
102
+ ]
103
+ },
104
+ "observation.images.rgb.head": {
105
+ "type": "VISUAL",
106
+ "shape": [
107
+ 720,
108
+ 720,
109
+ 3
110
+ ]
111
+ },
112
+ "observation.images.depth.left_wrist": {
113
+ "type": "VISUAL",
114
+ "shape": [
115
+ 480,
116
+ 480,
117
+ 3
118
+ ]
119
+ },
120
+ "observation.images.depth.right_wrist": {
121
+ "type": "VISUAL",
122
+ "shape": [
123
+ 480,
124
+ 480,
125
+ 3
126
+ ]
127
+ },
128
+ "observation.images.depth.head": {
129
+ "type": "VISUAL",
130
+ "shape": [
131
+ 720,
132
+ 720,
133
+ 3
134
+ ]
135
+ },
136
+ "observation.images.seg_instance_id.left_wrist": {
137
+ "type": "VISUAL",
138
+ "shape": [
139
+ 480,
140
+ 480,
141
+ 3
142
+ ]
143
+ },
144
+ "observation.images.seg_instance_id.right_wrist": {
145
+ "type": "VISUAL",
146
+ "shape": [
147
+ 480,
148
+ 480,
149
+ 3
150
+ ]
151
+ },
152
+ "observation.images.seg_instance_id.head": {
153
+ "type": "VISUAL",
154
+ "shape": [
155
+ 720,
156
+ 720,
157
+ 3
158
+ ]
159
+ },
160
+ "observation.cam_rel_poses": {
161
+ "type": "STATE",
162
+ "shape": [
163
+ 21
164
+ ]
165
+ },
166
+ "observation.state": {
167
+ "type": "STATE",
168
+ "shape": [
169
+ 256
170
+ ]
171
+ },
172
+ "observation.task_info": {
173
+ "type": "STATE",
174
+ "shape": [
175
+ 46
176
+ ]
177
+ }
178
+ },
179
+ "output_features": {
180
+ "action": {
181
+ "type": "ACTION",
182
+ "shape": [
183
+ 23
184
+ ]
185
+ }
186
+ },
187
+ "device": "cuda",
188
+ "use_amp": true,
189
+ "push_to_hub": true,
190
+ "repo_id": "ywu67/b1k-task0000",
191
+ "private": null,
192
+ "tags": null,
193
+ "license": null,
194
+ "pretrained_path": null,
195
+ "chunk_size": 100,
196
+ "n_action_steps": 100,
197
+ "normalization_mapping": {
198
+ "VISUAL": "MEAN_STD",
199
+ "STATE": "MEAN_STD",
200
+ "ACTION": "MEAN_STD"
201
+ },
202
+ "vision_backbone": "resnet18",
203
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
204
+ "replace_final_stride_with_dilation": false,
205
+ "pre_norm": false,
206
+ "dim_model": 512,
207
+ "n_heads": 8,
208
+ "dim_feedforward": 3200,
209
+ "feedforward_activation": "relu",
210
+ "n_encoder_layers": 4,
211
+ "n_decoder_layers": 1,
212
+ "use_vae": true,
213
+ "latent_dim": 32,
214
+ "n_vae_encoder_layers": 4,
215
+ "temporal_ensemble_coeff": null,
216
+ "dropout": 0.1,
217
+ "kl_weight": 10.0,
218
+ "optimizer_lr": 1e-05,
219
+ "optimizer_weight_decay": 0.0001,
220
+ "optimizer_lr_backbone": 1e-05
221
+ },
222
+ "output_dir": "outputs/task_0000_training",
223
+ "job_name": "act",
224
+ "resume": false,
225
+ "seed": 1000,
226
+ "num_workers": 12,
227
+ "batch_size": 16,
228
+ "steps": 100000,
229
+ "eval_freq": 20000,
230
+ "log_freq": 200,
231
+ "save_checkpoint": true,
232
+ "save_freq": 5000,
233
+ "use_policy_training_preset": true,
234
+ "optimizer": {
235
+ "type": "adamw",
236
+ "lr": 1e-05,
237
+ "weight_decay": 0.0001,
238
+ "grad_clip_norm": 10.0,
239
+ "betas": [
240
+ 0.9,
241
+ 0.999
242
+ ],
243
+ "eps": 1e-08
244
+ },
245
+ "scheduler": null,
246
+ "eval": {
247
+ "n_episodes": 50,
248
+ "batch_size": 50,
249
+ "use_async_envs": false
250
+ },
251
+ "wandb": {
252
+ "enable": true,
253
+ "disable_artifact": false,
254
+ "project": "lerobot-2025-challenge",
255
+ "entity": null,
256
+ "notes": "Training with torchcodec, batch_size=16, num_workers=12",
257
+ "run_id": "130gsib5",
258
+ "mode": null
259
+ },
260
+ "checkpoint_path": null,
261
+ "rename_map": {}
262
+ }