Sendera commited on
Commit
022db6d
·
verified ·
1 Parent(s): 96410c1

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +1 -0
  2. checkpoints/0100000/pretrained_model/config.json +103 -0
  3. checkpoints/0100000/pretrained_model/model.safetensors +3 -0
  4. checkpoints/0100000/pretrained_model/policy_postprocessor.json +32 -0
  5. checkpoints/0100000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  6. checkpoints/0100000/pretrained_model/policy_preprocessor.json +96 -0
  7. checkpoints/0100000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  8. checkpoints/0100000/pretrained_model/train_config.json +258 -0
  9. checkpoints/0100000/training_state/optimizer_param_groups.json +191 -0
  10. checkpoints/0100000/training_state/optimizer_state.safetensors +3 -0
  11. checkpoints/0100000/training_state/rng_state.safetensors +3 -0
  12. checkpoints/0100000/training_state/training_step.json +3 -0
  13. checkpoints/0200000/pretrained_model/config.json +103 -0
  14. checkpoints/0200000/pretrained_model/model.safetensors +3 -0
  15. checkpoints/0200000/pretrained_model/policy_postprocessor.json +32 -0
  16. checkpoints/0200000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  17. checkpoints/0200000/pretrained_model/policy_preprocessor.json +96 -0
  18. checkpoints/0200000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  19. checkpoints/0200000/pretrained_model/train_config.json +258 -0
  20. checkpoints/0200000/training_state/optimizer_param_groups.json +191 -0
  21. checkpoints/0200000/training_state/optimizer_state.safetensors +3 -0
  22. checkpoints/0200000/training_state/rng_state.safetensors +3 -0
  23. checkpoints/0200000/training_state/training_step.json +3 -0
  24. checkpoints/0300000/pretrained_model/config.json +103 -0
  25. checkpoints/0300000/pretrained_model/model.safetensors +3 -0
  26. checkpoints/0300000/pretrained_model/policy_postprocessor.json +32 -0
  27. checkpoints/0300000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  28. checkpoints/0300000/pretrained_model/policy_preprocessor.json +96 -0
  29. checkpoints/0300000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  30. checkpoints/0300000/pretrained_model/train_config.json +258 -0
  31. checkpoints/0300000/training_state/optimizer_param_groups.json +191 -0
  32. checkpoints/0300000/training_state/optimizer_state.safetensors +3 -0
  33. checkpoints/0300000/training_state/rng_state.safetensors +3 -0
  34. checkpoints/0300000/training_state/training_step.json +3 -0
  35. checkpoints/0400000/pretrained_model/config.json +103 -0
  36. checkpoints/0400000/pretrained_model/model.safetensors +3 -0
  37. checkpoints/0400000/pretrained_model/policy_postprocessor.json +32 -0
  38. checkpoints/0400000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  39. checkpoints/0400000/pretrained_model/policy_preprocessor.json +96 -0
  40. checkpoints/0400000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  41. checkpoints/0400000/pretrained_model/train_config.json +258 -0
  42. checkpoints/0400000/training_state/optimizer_param_groups.json +191 -0
  43. checkpoints/0400000/training_state/optimizer_state.safetensors +3 -0
  44. checkpoints/0400000/training_state/rng_state.safetensors +3 -0
  45. checkpoints/0400000/training_state/training_step.json +3 -0
  46. wandb/debug-internal.log +6 -0
  47. wandb/debug.log +19 -0
  48. wandb/run-20260211_163721-efad68xu/files/output.log +0 -0
  49. wandb/run-20260211_163721-efad68xu/files/requirements.txt +123 -0
  50. wandb/run-20260211_163721-efad68xu/files/wandb-metadata.json +64 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ wandb/run-20260211_163721-efad68xu/run-efad68xu.wandb filter=lfs diff=lfs merge=lfs -text
checkpoints/0100000/pretrained_model/config.json ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "act",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 14
9
+ ]
10
+ },
11
+ "observation.images.Cam1_global": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 256,
16
+ 256
17
+ ]
18
+ },
19
+ "observation.images.Cam1_local": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 256,
24
+ 256
25
+ ]
26
+ },
27
+ "observation.images.Cam1_red": {
28
+ "type": "VISUAL",
29
+ "shape": [
30
+ 3,
31
+ 256,
32
+ 256
33
+ ]
34
+ },
35
+ "observation.images.Cam2_global": {
36
+ "type": "VISUAL",
37
+ "shape": [
38
+ 3,
39
+ 256,
40
+ 256
41
+ ]
42
+ },
43
+ "observation.images.Cam2_local": {
44
+ "type": "VISUAL",
45
+ "shape": [
46
+ 3,
47
+ 256,
48
+ 256
49
+ ]
50
+ },
51
+ "observation.images.Cam2_red": {
52
+ "type": "VISUAL",
53
+ "shape": [
54
+ 3,
55
+ 256,
56
+ 256
57
+ ]
58
+ }
59
+ },
60
+ "output_features": {
61
+ "action": {
62
+ "type": "ACTION",
63
+ "shape": [
64
+ 6
65
+ ]
66
+ }
67
+ },
68
+ "device": "cuda",
69
+ "use_amp": false,
70
+ "use_peft": false,
71
+ "push_to_hub": false,
72
+ "repo_id": "act_pick5_drq",
73
+ "private": null,
74
+ "tags": null,
75
+ "license": null,
76
+ "pretrained_path": null,
77
+ "chunk_size": 100,
78
+ "n_action_steps": 100,
79
+ "normalization_mapping": {
80
+ "VISUAL": "MEAN_STD",
81
+ "STATE": "MEAN_STD",
82
+ "ACTION": "MEAN_STD"
83
+ },
84
+ "vision_backbone": "resnet18",
85
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
86
+ "replace_final_stride_with_dilation": false,
87
+ "pre_norm": false,
88
+ "dim_model": 512,
89
+ "n_heads": 8,
90
+ "dim_feedforward": 3200,
91
+ "feedforward_activation": "relu",
92
+ "n_encoder_layers": 4,
93
+ "n_decoder_layers": 1,
94
+ "use_vae": true,
95
+ "latent_dim": 32,
96
+ "n_vae_encoder_layers": 4,
97
+ "temporal_ensemble_coeff": null,
98
+ "dropout": 0.1,
99
+ "kl_weight": 10.0,
100
+ "optimizer_lr": 1e-05,
101
+ "optimizer_weight_decay": 0.0001,
102
+ "optimizer_lr_backbone": 1e-05
103
+ }
checkpoints/0100000/pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cdc77c3e04fa42079b203a8da99ad93544c7d0168a3d8bc11af2594a9e52c0c1
3
+ size 206732504
checkpoints/0100000/pretrained_model/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
+ 6
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
+ }
checkpoints/0100000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:478abd144078223b8649c694bd8f997c3019eb2fc501c05c99d3cb0cc6130745
3
+ size 12160
checkpoints/0100000/pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ 14
30
+ ]
31
+ },
32
+ "observation.images.Cam1_global": {
33
+ "type": "VISUAL",
34
+ "shape": [
35
+ 3,
36
+ 256,
37
+ 256
38
+ ]
39
+ },
40
+ "observation.images.Cam1_local": {
41
+ "type": "VISUAL",
42
+ "shape": [
43
+ 3,
44
+ 256,
45
+ 256
46
+ ]
47
+ },
48
+ "observation.images.Cam1_red": {
49
+ "type": "VISUAL",
50
+ "shape": [
51
+ 3,
52
+ 256,
53
+ 256
54
+ ]
55
+ },
56
+ "observation.images.Cam2_global": {
57
+ "type": "VISUAL",
58
+ "shape": [
59
+ 3,
60
+ 256,
61
+ 256
62
+ ]
63
+ },
64
+ "observation.images.Cam2_local": {
65
+ "type": "VISUAL",
66
+ "shape": [
67
+ 3,
68
+ 256,
69
+ 256
70
+ ]
71
+ },
72
+ "observation.images.Cam2_red": {
73
+ "type": "VISUAL",
74
+ "shape": [
75
+ 3,
76
+ 256,
77
+ 256
78
+ ]
79
+ },
80
+ "action": {
81
+ "type": "ACTION",
82
+ "shape": [
83
+ 6
84
+ ]
85
+ }
86
+ },
87
+ "norm_map": {
88
+ "VISUAL": "MEAN_STD",
89
+ "STATE": "MEAN_STD",
90
+ "ACTION": "MEAN_STD"
91
+ }
92
+ },
93
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
94
+ }
95
+ ]
96
+ }
checkpoints/0100000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:478abd144078223b8649c694bd8f997c3019eb2fc501c05c99d3cb0cc6130745
3
+ size 12160
checkpoints/0100000/pretrained_model/train_config.json ADDED
@@ -0,0 +1,258 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "Sendera/pick5",
4
+ "root": "/srv/scratch/z5649701/datasets/Sendera/pick5",
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": true,
8
+ "max_num_transforms": 3,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "brightness": {
12
+ "weight": 0.0,
13
+ "type": "ColorJitter",
14
+ "kwargs": {
15
+ "brightness": [
16
+ 0.8,
17
+ 1.2
18
+ ]
19
+ }
20
+ },
21
+ "contrast": {
22
+ "weight": 0.0,
23
+ "type": "ColorJitter",
24
+ "kwargs": {
25
+ "contrast": [
26
+ 0.8,
27
+ 1.2
28
+ ]
29
+ }
30
+ },
31
+ "saturation": {
32
+ "weight": 0.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "saturation": [
36
+ 0.5,
37
+ 1.5
38
+ ]
39
+ }
40
+ },
41
+ "hue": {
42
+ "weight": 0.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "hue": [
46
+ -0.05,
47
+ 0.05
48
+ ]
49
+ }
50
+ },
51
+ "sharpness": {
52
+ "weight": 0.0,
53
+ "type": "SharpnessJitter",
54
+ "kwargs": {
55
+ "sharpness": [
56
+ 0.5,
57
+ 1.5
58
+ ]
59
+ }
60
+ },
61
+ "affine": {
62
+ "weight": 0.0,
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
+ "resize": {
76
+ "weight": 0.0,
77
+ "type": "Resize",
78
+ "kwargs": {
79
+ "size": 512,
80
+ "interpolation": 2,
81
+ "antialias": true
82
+ }
83
+ },
84
+ "center_crop": {
85
+ "weight": 0.0,
86
+ "type": "CenterCrop",
87
+ "kwargs": {
88
+ "size": 512
89
+ }
90
+ },
91
+ "random_crop": {
92
+ "weight": 1.0,
93
+ "type": "RandomCrop",
94
+ "kwargs": {
95
+ "size": 256,
96
+ "padding": 13,
97
+ "padding_mode": "edge"
98
+ }
99
+ }
100
+ }
101
+ },
102
+ "revision": null,
103
+ "use_imagenet_stats": true,
104
+ "video_backend": "pyav",
105
+ "streaming": false
106
+ },
107
+ "env": null,
108
+ "policy": {
109
+ "type": "act",
110
+ "n_obs_steps": 1,
111
+ "input_features": {
112
+ "observation.state": {
113
+ "type": "STATE",
114
+ "shape": [
115
+ 14
116
+ ]
117
+ },
118
+ "observation.images.Cam1_global": {
119
+ "type": "VISUAL",
120
+ "shape": [
121
+ 3,
122
+ 256,
123
+ 256
124
+ ]
125
+ },
126
+ "observation.images.Cam1_local": {
127
+ "type": "VISUAL",
128
+ "shape": [
129
+ 3,
130
+ 256,
131
+ 256
132
+ ]
133
+ },
134
+ "observation.images.Cam1_red": {
135
+ "type": "VISUAL",
136
+ "shape": [
137
+ 3,
138
+ 256,
139
+ 256
140
+ ]
141
+ },
142
+ "observation.images.Cam2_global": {
143
+ "type": "VISUAL",
144
+ "shape": [
145
+ 3,
146
+ 256,
147
+ 256
148
+ ]
149
+ },
150
+ "observation.images.Cam2_local": {
151
+ "type": "VISUAL",
152
+ "shape": [
153
+ 3,
154
+ 256,
155
+ 256
156
+ ]
157
+ },
158
+ "observation.images.Cam2_red": {
159
+ "type": "VISUAL",
160
+ "shape": [
161
+ 3,
162
+ 256,
163
+ 256
164
+ ]
165
+ }
166
+ },
167
+ "output_features": {
168
+ "action": {
169
+ "type": "ACTION",
170
+ "shape": [
171
+ 6
172
+ ]
173
+ }
174
+ },
175
+ "device": "cuda",
176
+ "use_amp": false,
177
+ "use_peft": false,
178
+ "push_to_hub": false,
179
+ "repo_id": "act_pick5_drq",
180
+ "private": null,
181
+ "tags": null,
182
+ "license": null,
183
+ "pretrained_path": null,
184
+ "chunk_size": 100,
185
+ "n_action_steps": 100,
186
+ "normalization_mapping": {
187
+ "VISUAL": "MEAN_STD",
188
+ "STATE": "MEAN_STD",
189
+ "ACTION": "MEAN_STD"
190
+ },
191
+ "vision_backbone": "resnet18",
192
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
193
+ "replace_final_stride_with_dilation": false,
194
+ "pre_norm": false,
195
+ "dim_model": 512,
196
+ "n_heads": 8,
197
+ "dim_feedforward": 3200,
198
+ "feedforward_activation": "relu",
199
+ "n_encoder_layers": 4,
200
+ "n_decoder_layers": 1,
201
+ "use_vae": true,
202
+ "latent_dim": 32,
203
+ "n_vae_encoder_layers": 4,
204
+ "temporal_ensemble_coeff": null,
205
+ "dropout": 0.1,
206
+ "kl_weight": 10.0,
207
+ "optimizer_lr": 1e-05,
208
+ "optimizer_weight_decay": 0.0001,
209
+ "optimizer_lr_backbone": 1e-05
210
+ },
211
+ "output_dir": "outputs/train/2026-02-11/16-37-18_act_pick5_drq",
212
+ "job_name": "act_pick5_drq",
213
+ "resume": false,
214
+ "seed": 1000,
215
+ "num_workers": 12,
216
+ "batch_size": 8,
217
+ "steps": 1000000,
218
+ "eval_freq": 20000,
219
+ "log_freq": 200,
220
+ "tolerance_s": 0.0001,
221
+ "save_checkpoint": true,
222
+ "save_freq": 100000,
223
+ "use_policy_training_preset": true,
224
+ "optimizer": {
225
+ "type": "adamw",
226
+ "lr": 1e-05,
227
+ "weight_decay": 0.0001,
228
+ "grad_clip_norm": 10.0,
229
+ "betas": [
230
+ 0.9,
231
+ 0.999
232
+ ],
233
+ "eps": 1e-08
234
+ },
235
+ "scheduler": null,
236
+ "eval": {
237
+ "n_episodes": 50,
238
+ "batch_size": 50,
239
+ "use_async_envs": false
240
+ },
241
+ "wandb": {
242
+ "enable": true,
243
+ "disable_artifact": false,
244
+ "project": "lerobot_leo",
245
+ "entity": null,
246
+ "notes": null,
247
+ "run_id": "efad68xu",
248
+ "mode": null
249
+ },
250
+ "peft": null,
251
+ "use_rabc": false,
252
+ "rabc_progress_path": null,
253
+ "rabc_kappa": 0.01,
254
+ "rabc_epsilon": 1e-06,
255
+ "rabc_head_mode": "sparse",
256
+ "rename_map": {},
257
+ "checkpoint_path": null
258
+ }
checkpoints/0100000/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,191 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 1e-05,
4
+ "betas": [
5
+ 0.9,
6
+ 0.999
7
+ ],
8
+ "eps": 1e-08,
9
+ "weight_decay": 0.0001,
10
+ "amsgrad": false,
11
+ "maximize": false,
12
+ "foreach": null,
13
+ "capturable": false,
14
+ "differentiable": false,
15
+ "fused": null,
16
+ "decoupled_weight_decay": true,
17
+ "params": [
18
+ 0,
19
+ 1,
20
+ 2,
21
+ 3,
22
+ 4,
23
+ 5,
24
+ 6,
25
+ 7,
26
+ 8,
27
+ 9,
28
+ 10,
29
+ 11,
30
+ 12,
31
+ 13,
32
+ 14,
33
+ 15,
34
+ 16,
35
+ 17,
36
+ 18,
37
+ 19,
38
+ 20,
39
+ 21,
40
+ 22,
41
+ 23,
42
+ 24,
43
+ 25,
44
+ 26,
45
+ 27,
46
+ 28,
47
+ 29,
48
+ 30,
49
+ 31,
50
+ 32,
51
+ 33,
52
+ 34,
53
+ 35,
54
+ 36,
55
+ 37,
56
+ 38,
57
+ 39,
58
+ 40,
59
+ 41,
60
+ 42,
61
+ 43,
62
+ 44,
63
+ 45,
64
+ 46,
65
+ 47,
66
+ 48,
67
+ 49,
68
+ 50,
69
+ 51,
70
+ 52,
71
+ 53,
72
+ 54,
73
+ 55,
74
+ 56,
75
+ 57,
76
+ 58,
77
+ 59,
78
+ 60,
79
+ 61,
80
+ 62,
81
+ 63,
82
+ 64,
83
+ 65,
84
+ 66,
85
+ 67,
86
+ 68,
87
+ 69,
88
+ 70,
89
+ 71,
90
+ 72,
91
+ 73,
92
+ 74,
93
+ 75,
94
+ 76,
95
+ 77,
96
+ 78,
97
+ 79,
98
+ 80,
99
+ 81,
100
+ 82,
101
+ 83,
102
+ 84,
103
+ 85,
104
+ 86,
105
+ 87,
106
+ 88,
107
+ 89,
108
+ 90,
109
+ 91,
110
+ 92,
111
+ 93,
112
+ 94,
113
+ 95,
114
+ 96,
115
+ 97,
116
+ 98,
117
+ 99,
118
+ 100,
119
+ 101,
120
+ 102,
121
+ 103,
122
+ 104,
123
+ 105,
124
+ 106,
125
+ 107,
126
+ 108,
127
+ 109,
128
+ 110,
129
+ 111,
130
+ 112,
131
+ 113,
132
+ 114,
133
+ 115,
134
+ 116,
135
+ 117,
136
+ 118,
137
+ 119,
138
+ 120,
139
+ 121,
140
+ 122,
141
+ 123,
142
+ 124,
143
+ 125,
144
+ 126,
145
+ 127,
146
+ 128,
147
+ 129,
148
+ 130,
149
+ 131,
150
+ 132
151
+ ]
152
+ },
153
+ {
154
+ "lr": 1e-05,
155
+ "betas": [
156
+ 0.9,
157
+ 0.999
158
+ ],
159
+ "eps": 1e-08,
160
+ "weight_decay": 0.0001,
161
+ "amsgrad": false,
162
+ "maximize": false,
163
+ "foreach": null,
164
+ "capturable": false,
165
+ "differentiable": false,
166
+ "fused": null,
167
+ "decoupled_weight_decay": true,
168
+ "params": [
169
+ 133,
170
+ 134,
171
+ 135,
172
+ 136,
173
+ 137,
174
+ 138,
175
+ 139,
176
+ 140,
177
+ 141,
178
+ 142,
179
+ 143,
180
+ 144,
181
+ 145,
182
+ 146,
183
+ 147,
184
+ 148,
185
+ 149,
186
+ 150,
187
+ 151,
188
+ 152
189
+ ]
190
+ }
191
+ ]
checkpoints/0100000/training_state/optimizer_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3630cc20d53a352bdc69c962f0ce7af2c6d6cad9cb909306205c2f03f26675c
3
+ size 412883196
checkpoints/0100000/training_state/rng_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d919f01482ed7efc9af7ec7b617c76ad79745934eaf9a0498410484bdabd494
3
+ size 15708
checkpoints/0100000/training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 100000
3
+ }
checkpoints/0200000/pretrained_model/config.json ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "act",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 14
9
+ ]
10
+ },
11
+ "observation.images.Cam1_global": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 256,
16
+ 256
17
+ ]
18
+ },
19
+ "observation.images.Cam1_local": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 256,
24
+ 256
25
+ ]
26
+ },
27
+ "observation.images.Cam1_red": {
28
+ "type": "VISUAL",
29
+ "shape": [
30
+ 3,
31
+ 256,
32
+ 256
33
+ ]
34
+ },
35
+ "observation.images.Cam2_global": {
36
+ "type": "VISUAL",
37
+ "shape": [
38
+ 3,
39
+ 256,
40
+ 256
41
+ ]
42
+ },
43
+ "observation.images.Cam2_local": {
44
+ "type": "VISUAL",
45
+ "shape": [
46
+ 3,
47
+ 256,
48
+ 256
49
+ ]
50
+ },
51
+ "observation.images.Cam2_red": {
52
+ "type": "VISUAL",
53
+ "shape": [
54
+ 3,
55
+ 256,
56
+ 256
57
+ ]
58
+ }
59
+ },
60
+ "output_features": {
61
+ "action": {
62
+ "type": "ACTION",
63
+ "shape": [
64
+ 6
65
+ ]
66
+ }
67
+ },
68
+ "device": "cuda",
69
+ "use_amp": false,
70
+ "use_peft": false,
71
+ "push_to_hub": false,
72
+ "repo_id": "act_pick5_drq",
73
+ "private": null,
74
+ "tags": null,
75
+ "license": null,
76
+ "pretrained_path": null,
77
+ "chunk_size": 100,
78
+ "n_action_steps": 100,
79
+ "normalization_mapping": {
80
+ "VISUAL": "MEAN_STD",
81
+ "STATE": "MEAN_STD",
82
+ "ACTION": "MEAN_STD"
83
+ },
84
+ "vision_backbone": "resnet18",
85
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
86
+ "replace_final_stride_with_dilation": false,
87
+ "pre_norm": false,
88
+ "dim_model": 512,
89
+ "n_heads": 8,
90
+ "dim_feedforward": 3200,
91
+ "feedforward_activation": "relu",
92
+ "n_encoder_layers": 4,
93
+ "n_decoder_layers": 1,
94
+ "use_vae": true,
95
+ "latent_dim": 32,
96
+ "n_vae_encoder_layers": 4,
97
+ "temporal_ensemble_coeff": null,
98
+ "dropout": 0.1,
99
+ "kl_weight": 10.0,
100
+ "optimizer_lr": 1e-05,
101
+ "optimizer_weight_decay": 0.0001,
102
+ "optimizer_lr_backbone": 1e-05
103
+ }
checkpoints/0200000/pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6b6981476b1a4b37188d65b1933576407665fc3c6cae1caf6af8a88fd96f011b
3
+ size 206732504
checkpoints/0200000/pretrained_model/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
+ 6
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
+ }
checkpoints/0200000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:478abd144078223b8649c694bd8f997c3019eb2fc501c05c99d3cb0cc6130745
3
+ size 12160
checkpoints/0200000/pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ 14
30
+ ]
31
+ },
32
+ "observation.images.Cam1_global": {
33
+ "type": "VISUAL",
34
+ "shape": [
35
+ 3,
36
+ 256,
37
+ 256
38
+ ]
39
+ },
40
+ "observation.images.Cam1_local": {
41
+ "type": "VISUAL",
42
+ "shape": [
43
+ 3,
44
+ 256,
45
+ 256
46
+ ]
47
+ },
48
+ "observation.images.Cam1_red": {
49
+ "type": "VISUAL",
50
+ "shape": [
51
+ 3,
52
+ 256,
53
+ 256
54
+ ]
55
+ },
56
+ "observation.images.Cam2_global": {
57
+ "type": "VISUAL",
58
+ "shape": [
59
+ 3,
60
+ 256,
61
+ 256
62
+ ]
63
+ },
64
+ "observation.images.Cam2_local": {
65
+ "type": "VISUAL",
66
+ "shape": [
67
+ 3,
68
+ 256,
69
+ 256
70
+ ]
71
+ },
72
+ "observation.images.Cam2_red": {
73
+ "type": "VISUAL",
74
+ "shape": [
75
+ 3,
76
+ 256,
77
+ 256
78
+ ]
79
+ },
80
+ "action": {
81
+ "type": "ACTION",
82
+ "shape": [
83
+ 6
84
+ ]
85
+ }
86
+ },
87
+ "norm_map": {
88
+ "VISUAL": "MEAN_STD",
89
+ "STATE": "MEAN_STD",
90
+ "ACTION": "MEAN_STD"
91
+ }
92
+ },
93
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
94
+ }
95
+ ]
96
+ }
checkpoints/0200000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:478abd144078223b8649c694bd8f997c3019eb2fc501c05c99d3cb0cc6130745
3
+ size 12160
checkpoints/0200000/pretrained_model/train_config.json ADDED
@@ -0,0 +1,258 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "Sendera/pick5",
4
+ "root": "/srv/scratch/z5649701/datasets/Sendera/pick5",
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": true,
8
+ "max_num_transforms": 3,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "brightness": {
12
+ "weight": 0.0,
13
+ "type": "ColorJitter",
14
+ "kwargs": {
15
+ "brightness": [
16
+ 0.8,
17
+ 1.2
18
+ ]
19
+ }
20
+ },
21
+ "contrast": {
22
+ "weight": 0.0,
23
+ "type": "ColorJitter",
24
+ "kwargs": {
25
+ "contrast": [
26
+ 0.8,
27
+ 1.2
28
+ ]
29
+ }
30
+ },
31
+ "saturation": {
32
+ "weight": 0.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "saturation": [
36
+ 0.5,
37
+ 1.5
38
+ ]
39
+ }
40
+ },
41
+ "hue": {
42
+ "weight": 0.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "hue": [
46
+ -0.05,
47
+ 0.05
48
+ ]
49
+ }
50
+ },
51
+ "sharpness": {
52
+ "weight": 0.0,
53
+ "type": "SharpnessJitter",
54
+ "kwargs": {
55
+ "sharpness": [
56
+ 0.5,
57
+ 1.5
58
+ ]
59
+ }
60
+ },
61
+ "affine": {
62
+ "weight": 0.0,
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
+ "resize": {
76
+ "weight": 0.0,
77
+ "type": "Resize",
78
+ "kwargs": {
79
+ "size": 512,
80
+ "interpolation": 2,
81
+ "antialias": true
82
+ }
83
+ },
84
+ "center_crop": {
85
+ "weight": 0.0,
86
+ "type": "CenterCrop",
87
+ "kwargs": {
88
+ "size": 512
89
+ }
90
+ },
91
+ "random_crop": {
92
+ "weight": 1.0,
93
+ "type": "RandomCrop",
94
+ "kwargs": {
95
+ "size": 256,
96
+ "padding": 13,
97
+ "padding_mode": "edge"
98
+ }
99
+ }
100
+ }
101
+ },
102
+ "revision": null,
103
+ "use_imagenet_stats": true,
104
+ "video_backend": "pyav",
105
+ "streaming": false
106
+ },
107
+ "env": null,
108
+ "policy": {
109
+ "type": "act",
110
+ "n_obs_steps": 1,
111
+ "input_features": {
112
+ "observation.state": {
113
+ "type": "STATE",
114
+ "shape": [
115
+ 14
116
+ ]
117
+ },
118
+ "observation.images.Cam1_global": {
119
+ "type": "VISUAL",
120
+ "shape": [
121
+ 3,
122
+ 256,
123
+ 256
124
+ ]
125
+ },
126
+ "observation.images.Cam1_local": {
127
+ "type": "VISUAL",
128
+ "shape": [
129
+ 3,
130
+ 256,
131
+ 256
132
+ ]
133
+ },
134
+ "observation.images.Cam1_red": {
135
+ "type": "VISUAL",
136
+ "shape": [
137
+ 3,
138
+ 256,
139
+ 256
140
+ ]
141
+ },
142
+ "observation.images.Cam2_global": {
143
+ "type": "VISUAL",
144
+ "shape": [
145
+ 3,
146
+ 256,
147
+ 256
148
+ ]
149
+ },
150
+ "observation.images.Cam2_local": {
151
+ "type": "VISUAL",
152
+ "shape": [
153
+ 3,
154
+ 256,
155
+ 256
156
+ ]
157
+ },
158
+ "observation.images.Cam2_red": {
159
+ "type": "VISUAL",
160
+ "shape": [
161
+ 3,
162
+ 256,
163
+ 256
164
+ ]
165
+ }
166
+ },
167
+ "output_features": {
168
+ "action": {
169
+ "type": "ACTION",
170
+ "shape": [
171
+ 6
172
+ ]
173
+ }
174
+ },
175
+ "device": "cuda",
176
+ "use_amp": false,
177
+ "use_peft": false,
178
+ "push_to_hub": false,
179
+ "repo_id": "act_pick5_drq",
180
+ "private": null,
181
+ "tags": null,
182
+ "license": null,
183
+ "pretrained_path": null,
184
+ "chunk_size": 100,
185
+ "n_action_steps": 100,
186
+ "normalization_mapping": {
187
+ "VISUAL": "MEAN_STD",
188
+ "STATE": "MEAN_STD",
189
+ "ACTION": "MEAN_STD"
190
+ },
191
+ "vision_backbone": "resnet18",
192
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
193
+ "replace_final_stride_with_dilation": false,
194
+ "pre_norm": false,
195
+ "dim_model": 512,
196
+ "n_heads": 8,
197
+ "dim_feedforward": 3200,
198
+ "feedforward_activation": "relu",
199
+ "n_encoder_layers": 4,
200
+ "n_decoder_layers": 1,
201
+ "use_vae": true,
202
+ "latent_dim": 32,
203
+ "n_vae_encoder_layers": 4,
204
+ "temporal_ensemble_coeff": null,
205
+ "dropout": 0.1,
206
+ "kl_weight": 10.0,
207
+ "optimizer_lr": 1e-05,
208
+ "optimizer_weight_decay": 0.0001,
209
+ "optimizer_lr_backbone": 1e-05
210
+ },
211
+ "output_dir": "outputs/train/2026-02-11/16-37-18_act_pick5_drq",
212
+ "job_name": "act_pick5_drq",
213
+ "resume": false,
214
+ "seed": 1000,
215
+ "num_workers": 12,
216
+ "batch_size": 8,
217
+ "steps": 1000000,
218
+ "eval_freq": 20000,
219
+ "log_freq": 200,
220
+ "tolerance_s": 0.0001,
221
+ "save_checkpoint": true,
222
+ "save_freq": 100000,
223
+ "use_policy_training_preset": true,
224
+ "optimizer": {
225
+ "type": "adamw",
226
+ "lr": 1e-05,
227
+ "weight_decay": 0.0001,
228
+ "grad_clip_norm": 10.0,
229
+ "betas": [
230
+ 0.9,
231
+ 0.999
232
+ ],
233
+ "eps": 1e-08
234
+ },
235
+ "scheduler": null,
236
+ "eval": {
237
+ "n_episodes": 50,
238
+ "batch_size": 50,
239
+ "use_async_envs": false
240
+ },
241
+ "wandb": {
242
+ "enable": true,
243
+ "disable_artifact": false,
244
+ "project": "lerobot_leo",
245
+ "entity": null,
246
+ "notes": null,
247
+ "run_id": "efad68xu",
248
+ "mode": null
249
+ },
250
+ "peft": null,
251
+ "use_rabc": false,
252
+ "rabc_progress_path": null,
253
+ "rabc_kappa": 0.01,
254
+ "rabc_epsilon": 1e-06,
255
+ "rabc_head_mode": "sparse",
256
+ "rename_map": {},
257
+ "checkpoint_path": null
258
+ }
checkpoints/0200000/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,191 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 1e-05,
4
+ "betas": [
5
+ 0.9,
6
+ 0.999
7
+ ],
8
+ "eps": 1e-08,
9
+ "weight_decay": 0.0001,
10
+ "amsgrad": false,
11
+ "maximize": false,
12
+ "foreach": null,
13
+ "capturable": false,
14
+ "differentiable": false,
15
+ "fused": null,
16
+ "decoupled_weight_decay": true,
17
+ "params": [
18
+ 0,
19
+ 1,
20
+ 2,
21
+ 3,
22
+ 4,
23
+ 5,
24
+ 6,
25
+ 7,
26
+ 8,
27
+ 9,
28
+ 10,
29
+ 11,
30
+ 12,
31
+ 13,
32
+ 14,
33
+ 15,
34
+ 16,
35
+ 17,
36
+ 18,
37
+ 19,
38
+ 20,
39
+ 21,
40
+ 22,
41
+ 23,
42
+ 24,
43
+ 25,
44
+ 26,
45
+ 27,
46
+ 28,
47
+ 29,
48
+ 30,
49
+ 31,
50
+ 32,
51
+ 33,
52
+ 34,
53
+ 35,
54
+ 36,
55
+ 37,
56
+ 38,
57
+ 39,
58
+ 40,
59
+ 41,
60
+ 42,
61
+ 43,
62
+ 44,
63
+ 45,
64
+ 46,
65
+ 47,
66
+ 48,
67
+ 49,
68
+ 50,
69
+ 51,
70
+ 52,
71
+ 53,
72
+ 54,
73
+ 55,
74
+ 56,
75
+ 57,
76
+ 58,
77
+ 59,
78
+ 60,
79
+ 61,
80
+ 62,
81
+ 63,
82
+ 64,
83
+ 65,
84
+ 66,
85
+ 67,
86
+ 68,
87
+ 69,
88
+ 70,
89
+ 71,
90
+ 72,
91
+ 73,
92
+ 74,
93
+ 75,
94
+ 76,
95
+ 77,
96
+ 78,
97
+ 79,
98
+ 80,
99
+ 81,
100
+ 82,
101
+ 83,
102
+ 84,
103
+ 85,
104
+ 86,
105
+ 87,
106
+ 88,
107
+ 89,
108
+ 90,
109
+ 91,
110
+ 92,
111
+ 93,
112
+ 94,
113
+ 95,
114
+ 96,
115
+ 97,
116
+ 98,
117
+ 99,
118
+ 100,
119
+ 101,
120
+ 102,
121
+ 103,
122
+ 104,
123
+ 105,
124
+ 106,
125
+ 107,
126
+ 108,
127
+ 109,
128
+ 110,
129
+ 111,
130
+ 112,
131
+ 113,
132
+ 114,
133
+ 115,
134
+ 116,
135
+ 117,
136
+ 118,
137
+ 119,
138
+ 120,
139
+ 121,
140
+ 122,
141
+ 123,
142
+ 124,
143
+ 125,
144
+ 126,
145
+ 127,
146
+ 128,
147
+ 129,
148
+ 130,
149
+ 131,
150
+ 132
151
+ ]
152
+ },
153
+ {
154
+ "lr": 1e-05,
155
+ "betas": [
156
+ 0.9,
157
+ 0.999
158
+ ],
159
+ "eps": 1e-08,
160
+ "weight_decay": 0.0001,
161
+ "amsgrad": false,
162
+ "maximize": false,
163
+ "foreach": null,
164
+ "capturable": false,
165
+ "differentiable": false,
166
+ "fused": null,
167
+ "decoupled_weight_decay": true,
168
+ "params": [
169
+ 133,
170
+ 134,
171
+ 135,
172
+ 136,
173
+ 137,
174
+ 138,
175
+ 139,
176
+ 140,
177
+ 141,
178
+ 142,
179
+ 143,
180
+ 144,
181
+ 145,
182
+ 146,
183
+ 147,
184
+ 148,
185
+ 149,
186
+ 150,
187
+ 151,
188
+ 152
189
+ ]
190
+ }
191
+ ]
checkpoints/0200000/training_state/optimizer_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7fcdc266bb3efde42f0d83b60f891d94be11c7ecf2f8ee2343fdee4360985fc7
3
+ size 412883196
checkpoints/0200000/training_state/rng_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0790c37dc57c7df57d88b7e424292383e1de8a5d4a0e5d61ec06376202748925
3
+ size 15708
checkpoints/0200000/training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 200000
3
+ }
checkpoints/0300000/pretrained_model/config.json ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "act",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 14
9
+ ]
10
+ },
11
+ "observation.images.Cam1_global": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 256,
16
+ 256
17
+ ]
18
+ },
19
+ "observation.images.Cam1_local": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 256,
24
+ 256
25
+ ]
26
+ },
27
+ "observation.images.Cam1_red": {
28
+ "type": "VISUAL",
29
+ "shape": [
30
+ 3,
31
+ 256,
32
+ 256
33
+ ]
34
+ },
35
+ "observation.images.Cam2_global": {
36
+ "type": "VISUAL",
37
+ "shape": [
38
+ 3,
39
+ 256,
40
+ 256
41
+ ]
42
+ },
43
+ "observation.images.Cam2_local": {
44
+ "type": "VISUAL",
45
+ "shape": [
46
+ 3,
47
+ 256,
48
+ 256
49
+ ]
50
+ },
51
+ "observation.images.Cam2_red": {
52
+ "type": "VISUAL",
53
+ "shape": [
54
+ 3,
55
+ 256,
56
+ 256
57
+ ]
58
+ }
59
+ },
60
+ "output_features": {
61
+ "action": {
62
+ "type": "ACTION",
63
+ "shape": [
64
+ 6
65
+ ]
66
+ }
67
+ },
68
+ "device": "cuda",
69
+ "use_amp": false,
70
+ "use_peft": false,
71
+ "push_to_hub": false,
72
+ "repo_id": "act_pick5_drq",
73
+ "private": null,
74
+ "tags": null,
75
+ "license": null,
76
+ "pretrained_path": null,
77
+ "chunk_size": 100,
78
+ "n_action_steps": 100,
79
+ "normalization_mapping": {
80
+ "VISUAL": "MEAN_STD",
81
+ "STATE": "MEAN_STD",
82
+ "ACTION": "MEAN_STD"
83
+ },
84
+ "vision_backbone": "resnet18",
85
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
86
+ "replace_final_stride_with_dilation": false,
87
+ "pre_norm": false,
88
+ "dim_model": 512,
89
+ "n_heads": 8,
90
+ "dim_feedforward": 3200,
91
+ "feedforward_activation": "relu",
92
+ "n_encoder_layers": 4,
93
+ "n_decoder_layers": 1,
94
+ "use_vae": true,
95
+ "latent_dim": 32,
96
+ "n_vae_encoder_layers": 4,
97
+ "temporal_ensemble_coeff": null,
98
+ "dropout": 0.1,
99
+ "kl_weight": 10.0,
100
+ "optimizer_lr": 1e-05,
101
+ "optimizer_weight_decay": 0.0001,
102
+ "optimizer_lr_backbone": 1e-05
103
+ }
checkpoints/0300000/pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45ade94a151d2675f799dec71f13941a6d9a21f6f41bf075d804ccdba42e4303
3
+ size 206732504
checkpoints/0300000/pretrained_model/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
+ 6
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
+ }
checkpoints/0300000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:478abd144078223b8649c694bd8f997c3019eb2fc501c05c99d3cb0cc6130745
3
+ size 12160
checkpoints/0300000/pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ 14
30
+ ]
31
+ },
32
+ "observation.images.Cam1_global": {
33
+ "type": "VISUAL",
34
+ "shape": [
35
+ 3,
36
+ 256,
37
+ 256
38
+ ]
39
+ },
40
+ "observation.images.Cam1_local": {
41
+ "type": "VISUAL",
42
+ "shape": [
43
+ 3,
44
+ 256,
45
+ 256
46
+ ]
47
+ },
48
+ "observation.images.Cam1_red": {
49
+ "type": "VISUAL",
50
+ "shape": [
51
+ 3,
52
+ 256,
53
+ 256
54
+ ]
55
+ },
56
+ "observation.images.Cam2_global": {
57
+ "type": "VISUAL",
58
+ "shape": [
59
+ 3,
60
+ 256,
61
+ 256
62
+ ]
63
+ },
64
+ "observation.images.Cam2_local": {
65
+ "type": "VISUAL",
66
+ "shape": [
67
+ 3,
68
+ 256,
69
+ 256
70
+ ]
71
+ },
72
+ "observation.images.Cam2_red": {
73
+ "type": "VISUAL",
74
+ "shape": [
75
+ 3,
76
+ 256,
77
+ 256
78
+ ]
79
+ },
80
+ "action": {
81
+ "type": "ACTION",
82
+ "shape": [
83
+ 6
84
+ ]
85
+ }
86
+ },
87
+ "norm_map": {
88
+ "VISUAL": "MEAN_STD",
89
+ "STATE": "MEAN_STD",
90
+ "ACTION": "MEAN_STD"
91
+ }
92
+ },
93
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
94
+ }
95
+ ]
96
+ }
checkpoints/0300000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:478abd144078223b8649c694bd8f997c3019eb2fc501c05c99d3cb0cc6130745
3
+ size 12160
checkpoints/0300000/pretrained_model/train_config.json ADDED
@@ -0,0 +1,258 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "Sendera/pick5",
4
+ "root": "/srv/scratch/z5649701/datasets/Sendera/pick5",
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": true,
8
+ "max_num_transforms": 3,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "brightness": {
12
+ "weight": 0.0,
13
+ "type": "ColorJitter",
14
+ "kwargs": {
15
+ "brightness": [
16
+ 0.8,
17
+ 1.2
18
+ ]
19
+ }
20
+ },
21
+ "contrast": {
22
+ "weight": 0.0,
23
+ "type": "ColorJitter",
24
+ "kwargs": {
25
+ "contrast": [
26
+ 0.8,
27
+ 1.2
28
+ ]
29
+ }
30
+ },
31
+ "saturation": {
32
+ "weight": 0.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "saturation": [
36
+ 0.5,
37
+ 1.5
38
+ ]
39
+ }
40
+ },
41
+ "hue": {
42
+ "weight": 0.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "hue": [
46
+ -0.05,
47
+ 0.05
48
+ ]
49
+ }
50
+ },
51
+ "sharpness": {
52
+ "weight": 0.0,
53
+ "type": "SharpnessJitter",
54
+ "kwargs": {
55
+ "sharpness": [
56
+ 0.5,
57
+ 1.5
58
+ ]
59
+ }
60
+ },
61
+ "affine": {
62
+ "weight": 0.0,
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
+ "resize": {
76
+ "weight": 0.0,
77
+ "type": "Resize",
78
+ "kwargs": {
79
+ "size": 512,
80
+ "interpolation": 2,
81
+ "antialias": true
82
+ }
83
+ },
84
+ "center_crop": {
85
+ "weight": 0.0,
86
+ "type": "CenterCrop",
87
+ "kwargs": {
88
+ "size": 512
89
+ }
90
+ },
91
+ "random_crop": {
92
+ "weight": 1.0,
93
+ "type": "RandomCrop",
94
+ "kwargs": {
95
+ "size": 256,
96
+ "padding": 13,
97
+ "padding_mode": "edge"
98
+ }
99
+ }
100
+ }
101
+ },
102
+ "revision": null,
103
+ "use_imagenet_stats": true,
104
+ "video_backend": "pyav",
105
+ "streaming": false
106
+ },
107
+ "env": null,
108
+ "policy": {
109
+ "type": "act",
110
+ "n_obs_steps": 1,
111
+ "input_features": {
112
+ "observation.state": {
113
+ "type": "STATE",
114
+ "shape": [
115
+ 14
116
+ ]
117
+ },
118
+ "observation.images.Cam1_global": {
119
+ "type": "VISUAL",
120
+ "shape": [
121
+ 3,
122
+ 256,
123
+ 256
124
+ ]
125
+ },
126
+ "observation.images.Cam1_local": {
127
+ "type": "VISUAL",
128
+ "shape": [
129
+ 3,
130
+ 256,
131
+ 256
132
+ ]
133
+ },
134
+ "observation.images.Cam1_red": {
135
+ "type": "VISUAL",
136
+ "shape": [
137
+ 3,
138
+ 256,
139
+ 256
140
+ ]
141
+ },
142
+ "observation.images.Cam2_global": {
143
+ "type": "VISUAL",
144
+ "shape": [
145
+ 3,
146
+ 256,
147
+ 256
148
+ ]
149
+ },
150
+ "observation.images.Cam2_local": {
151
+ "type": "VISUAL",
152
+ "shape": [
153
+ 3,
154
+ 256,
155
+ 256
156
+ ]
157
+ },
158
+ "observation.images.Cam2_red": {
159
+ "type": "VISUAL",
160
+ "shape": [
161
+ 3,
162
+ 256,
163
+ 256
164
+ ]
165
+ }
166
+ },
167
+ "output_features": {
168
+ "action": {
169
+ "type": "ACTION",
170
+ "shape": [
171
+ 6
172
+ ]
173
+ }
174
+ },
175
+ "device": "cuda",
176
+ "use_amp": false,
177
+ "use_peft": false,
178
+ "push_to_hub": false,
179
+ "repo_id": "act_pick5_drq",
180
+ "private": null,
181
+ "tags": null,
182
+ "license": null,
183
+ "pretrained_path": null,
184
+ "chunk_size": 100,
185
+ "n_action_steps": 100,
186
+ "normalization_mapping": {
187
+ "VISUAL": "MEAN_STD",
188
+ "STATE": "MEAN_STD",
189
+ "ACTION": "MEAN_STD"
190
+ },
191
+ "vision_backbone": "resnet18",
192
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
193
+ "replace_final_stride_with_dilation": false,
194
+ "pre_norm": false,
195
+ "dim_model": 512,
196
+ "n_heads": 8,
197
+ "dim_feedforward": 3200,
198
+ "feedforward_activation": "relu",
199
+ "n_encoder_layers": 4,
200
+ "n_decoder_layers": 1,
201
+ "use_vae": true,
202
+ "latent_dim": 32,
203
+ "n_vae_encoder_layers": 4,
204
+ "temporal_ensemble_coeff": null,
205
+ "dropout": 0.1,
206
+ "kl_weight": 10.0,
207
+ "optimizer_lr": 1e-05,
208
+ "optimizer_weight_decay": 0.0001,
209
+ "optimizer_lr_backbone": 1e-05
210
+ },
211
+ "output_dir": "outputs/train/2026-02-11/16-37-18_act_pick5_drq",
212
+ "job_name": "act_pick5_drq",
213
+ "resume": false,
214
+ "seed": 1000,
215
+ "num_workers": 12,
216
+ "batch_size": 8,
217
+ "steps": 1000000,
218
+ "eval_freq": 20000,
219
+ "log_freq": 200,
220
+ "tolerance_s": 0.0001,
221
+ "save_checkpoint": true,
222
+ "save_freq": 100000,
223
+ "use_policy_training_preset": true,
224
+ "optimizer": {
225
+ "type": "adamw",
226
+ "lr": 1e-05,
227
+ "weight_decay": 0.0001,
228
+ "grad_clip_norm": 10.0,
229
+ "betas": [
230
+ 0.9,
231
+ 0.999
232
+ ],
233
+ "eps": 1e-08
234
+ },
235
+ "scheduler": null,
236
+ "eval": {
237
+ "n_episodes": 50,
238
+ "batch_size": 50,
239
+ "use_async_envs": false
240
+ },
241
+ "wandb": {
242
+ "enable": true,
243
+ "disable_artifact": false,
244
+ "project": "lerobot_leo",
245
+ "entity": null,
246
+ "notes": null,
247
+ "run_id": "efad68xu",
248
+ "mode": null
249
+ },
250
+ "peft": null,
251
+ "use_rabc": false,
252
+ "rabc_progress_path": null,
253
+ "rabc_kappa": 0.01,
254
+ "rabc_epsilon": 1e-06,
255
+ "rabc_head_mode": "sparse",
256
+ "rename_map": {},
257
+ "checkpoint_path": null
258
+ }
checkpoints/0300000/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,191 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 1e-05,
4
+ "betas": [
5
+ 0.9,
6
+ 0.999
7
+ ],
8
+ "eps": 1e-08,
9
+ "weight_decay": 0.0001,
10
+ "amsgrad": false,
11
+ "maximize": false,
12
+ "foreach": null,
13
+ "capturable": false,
14
+ "differentiable": false,
15
+ "fused": null,
16
+ "decoupled_weight_decay": true,
17
+ "params": [
18
+ 0,
19
+ 1,
20
+ 2,
21
+ 3,
22
+ 4,
23
+ 5,
24
+ 6,
25
+ 7,
26
+ 8,
27
+ 9,
28
+ 10,
29
+ 11,
30
+ 12,
31
+ 13,
32
+ 14,
33
+ 15,
34
+ 16,
35
+ 17,
36
+ 18,
37
+ 19,
38
+ 20,
39
+ 21,
40
+ 22,
41
+ 23,
42
+ 24,
43
+ 25,
44
+ 26,
45
+ 27,
46
+ 28,
47
+ 29,
48
+ 30,
49
+ 31,
50
+ 32,
51
+ 33,
52
+ 34,
53
+ 35,
54
+ 36,
55
+ 37,
56
+ 38,
57
+ 39,
58
+ 40,
59
+ 41,
60
+ 42,
61
+ 43,
62
+ 44,
63
+ 45,
64
+ 46,
65
+ 47,
66
+ 48,
67
+ 49,
68
+ 50,
69
+ 51,
70
+ 52,
71
+ 53,
72
+ 54,
73
+ 55,
74
+ 56,
75
+ 57,
76
+ 58,
77
+ 59,
78
+ 60,
79
+ 61,
80
+ 62,
81
+ 63,
82
+ 64,
83
+ 65,
84
+ 66,
85
+ 67,
86
+ 68,
87
+ 69,
88
+ 70,
89
+ 71,
90
+ 72,
91
+ 73,
92
+ 74,
93
+ 75,
94
+ 76,
95
+ 77,
96
+ 78,
97
+ 79,
98
+ 80,
99
+ 81,
100
+ 82,
101
+ 83,
102
+ 84,
103
+ 85,
104
+ 86,
105
+ 87,
106
+ 88,
107
+ 89,
108
+ 90,
109
+ 91,
110
+ 92,
111
+ 93,
112
+ 94,
113
+ 95,
114
+ 96,
115
+ 97,
116
+ 98,
117
+ 99,
118
+ 100,
119
+ 101,
120
+ 102,
121
+ 103,
122
+ 104,
123
+ 105,
124
+ 106,
125
+ 107,
126
+ 108,
127
+ 109,
128
+ 110,
129
+ 111,
130
+ 112,
131
+ 113,
132
+ 114,
133
+ 115,
134
+ 116,
135
+ 117,
136
+ 118,
137
+ 119,
138
+ 120,
139
+ 121,
140
+ 122,
141
+ 123,
142
+ 124,
143
+ 125,
144
+ 126,
145
+ 127,
146
+ 128,
147
+ 129,
148
+ 130,
149
+ 131,
150
+ 132
151
+ ]
152
+ },
153
+ {
154
+ "lr": 1e-05,
155
+ "betas": [
156
+ 0.9,
157
+ 0.999
158
+ ],
159
+ "eps": 1e-08,
160
+ "weight_decay": 0.0001,
161
+ "amsgrad": false,
162
+ "maximize": false,
163
+ "foreach": null,
164
+ "capturable": false,
165
+ "differentiable": false,
166
+ "fused": null,
167
+ "decoupled_weight_decay": true,
168
+ "params": [
169
+ 133,
170
+ 134,
171
+ 135,
172
+ 136,
173
+ 137,
174
+ 138,
175
+ 139,
176
+ 140,
177
+ 141,
178
+ 142,
179
+ 143,
180
+ 144,
181
+ 145,
182
+ 146,
183
+ 147,
184
+ 148,
185
+ 149,
186
+ 150,
187
+ 151,
188
+ 152
189
+ ]
190
+ }
191
+ ]
checkpoints/0300000/training_state/optimizer_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:18b83722b10d36196b3433efff091bb12ce320e35fdb83fc6171cc4c49de2e71
3
+ size 412883196
checkpoints/0300000/training_state/rng_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d058ab6a6b4a50f3200dacd13de35e9746de2c0f9db07cfde01abdffe2954083
3
+ size 15708
checkpoints/0300000/training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 300000
3
+ }
checkpoints/0400000/pretrained_model/config.json ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "act",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 14
9
+ ]
10
+ },
11
+ "observation.images.Cam1_global": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 256,
16
+ 256
17
+ ]
18
+ },
19
+ "observation.images.Cam1_local": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 256,
24
+ 256
25
+ ]
26
+ },
27
+ "observation.images.Cam1_red": {
28
+ "type": "VISUAL",
29
+ "shape": [
30
+ 3,
31
+ 256,
32
+ 256
33
+ ]
34
+ },
35
+ "observation.images.Cam2_global": {
36
+ "type": "VISUAL",
37
+ "shape": [
38
+ 3,
39
+ 256,
40
+ 256
41
+ ]
42
+ },
43
+ "observation.images.Cam2_local": {
44
+ "type": "VISUAL",
45
+ "shape": [
46
+ 3,
47
+ 256,
48
+ 256
49
+ ]
50
+ },
51
+ "observation.images.Cam2_red": {
52
+ "type": "VISUAL",
53
+ "shape": [
54
+ 3,
55
+ 256,
56
+ 256
57
+ ]
58
+ }
59
+ },
60
+ "output_features": {
61
+ "action": {
62
+ "type": "ACTION",
63
+ "shape": [
64
+ 6
65
+ ]
66
+ }
67
+ },
68
+ "device": "cuda",
69
+ "use_amp": false,
70
+ "use_peft": false,
71
+ "push_to_hub": false,
72
+ "repo_id": "act_pick5_drq",
73
+ "private": null,
74
+ "tags": null,
75
+ "license": null,
76
+ "pretrained_path": null,
77
+ "chunk_size": 100,
78
+ "n_action_steps": 100,
79
+ "normalization_mapping": {
80
+ "VISUAL": "MEAN_STD",
81
+ "STATE": "MEAN_STD",
82
+ "ACTION": "MEAN_STD"
83
+ },
84
+ "vision_backbone": "resnet18",
85
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
86
+ "replace_final_stride_with_dilation": false,
87
+ "pre_norm": false,
88
+ "dim_model": 512,
89
+ "n_heads": 8,
90
+ "dim_feedforward": 3200,
91
+ "feedforward_activation": "relu",
92
+ "n_encoder_layers": 4,
93
+ "n_decoder_layers": 1,
94
+ "use_vae": true,
95
+ "latent_dim": 32,
96
+ "n_vae_encoder_layers": 4,
97
+ "temporal_ensemble_coeff": null,
98
+ "dropout": 0.1,
99
+ "kl_weight": 10.0,
100
+ "optimizer_lr": 1e-05,
101
+ "optimizer_weight_decay": 0.0001,
102
+ "optimizer_lr_backbone": 1e-05
103
+ }
checkpoints/0400000/pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:910700b669c6f950958ff0a592312d1c447f76a7245524756140c88d1856acf6
3
+ size 206732504
checkpoints/0400000/pretrained_model/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
+ 6
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
+ }
checkpoints/0400000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:478abd144078223b8649c694bd8f997c3019eb2fc501c05c99d3cb0cc6130745
3
+ size 12160
checkpoints/0400000/pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ 14
30
+ ]
31
+ },
32
+ "observation.images.Cam1_global": {
33
+ "type": "VISUAL",
34
+ "shape": [
35
+ 3,
36
+ 256,
37
+ 256
38
+ ]
39
+ },
40
+ "observation.images.Cam1_local": {
41
+ "type": "VISUAL",
42
+ "shape": [
43
+ 3,
44
+ 256,
45
+ 256
46
+ ]
47
+ },
48
+ "observation.images.Cam1_red": {
49
+ "type": "VISUAL",
50
+ "shape": [
51
+ 3,
52
+ 256,
53
+ 256
54
+ ]
55
+ },
56
+ "observation.images.Cam2_global": {
57
+ "type": "VISUAL",
58
+ "shape": [
59
+ 3,
60
+ 256,
61
+ 256
62
+ ]
63
+ },
64
+ "observation.images.Cam2_local": {
65
+ "type": "VISUAL",
66
+ "shape": [
67
+ 3,
68
+ 256,
69
+ 256
70
+ ]
71
+ },
72
+ "observation.images.Cam2_red": {
73
+ "type": "VISUAL",
74
+ "shape": [
75
+ 3,
76
+ 256,
77
+ 256
78
+ ]
79
+ },
80
+ "action": {
81
+ "type": "ACTION",
82
+ "shape": [
83
+ 6
84
+ ]
85
+ }
86
+ },
87
+ "norm_map": {
88
+ "VISUAL": "MEAN_STD",
89
+ "STATE": "MEAN_STD",
90
+ "ACTION": "MEAN_STD"
91
+ }
92
+ },
93
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
94
+ }
95
+ ]
96
+ }
checkpoints/0400000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:478abd144078223b8649c694bd8f997c3019eb2fc501c05c99d3cb0cc6130745
3
+ size 12160
checkpoints/0400000/pretrained_model/train_config.json ADDED
@@ -0,0 +1,258 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "Sendera/pick5",
4
+ "root": "/srv/scratch/z5649701/datasets/Sendera/pick5",
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": true,
8
+ "max_num_transforms": 3,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "brightness": {
12
+ "weight": 0.0,
13
+ "type": "ColorJitter",
14
+ "kwargs": {
15
+ "brightness": [
16
+ 0.8,
17
+ 1.2
18
+ ]
19
+ }
20
+ },
21
+ "contrast": {
22
+ "weight": 0.0,
23
+ "type": "ColorJitter",
24
+ "kwargs": {
25
+ "contrast": [
26
+ 0.8,
27
+ 1.2
28
+ ]
29
+ }
30
+ },
31
+ "saturation": {
32
+ "weight": 0.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "saturation": [
36
+ 0.5,
37
+ 1.5
38
+ ]
39
+ }
40
+ },
41
+ "hue": {
42
+ "weight": 0.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "hue": [
46
+ -0.05,
47
+ 0.05
48
+ ]
49
+ }
50
+ },
51
+ "sharpness": {
52
+ "weight": 0.0,
53
+ "type": "SharpnessJitter",
54
+ "kwargs": {
55
+ "sharpness": [
56
+ 0.5,
57
+ 1.5
58
+ ]
59
+ }
60
+ },
61
+ "affine": {
62
+ "weight": 0.0,
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
+ "resize": {
76
+ "weight": 0.0,
77
+ "type": "Resize",
78
+ "kwargs": {
79
+ "size": 512,
80
+ "interpolation": 2,
81
+ "antialias": true
82
+ }
83
+ },
84
+ "center_crop": {
85
+ "weight": 0.0,
86
+ "type": "CenterCrop",
87
+ "kwargs": {
88
+ "size": 512
89
+ }
90
+ },
91
+ "random_crop": {
92
+ "weight": 1.0,
93
+ "type": "RandomCrop",
94
+ "kwargs": {
95
+ "size": 256,
96
+ "padding": 13,
97
+ "padding_mode": "edge"
98
+ }
99
+ }
100
+ }
101
+ },
102
+ "revision": null,
103
+ "use_imagenet_stats": true,
104
+ "video_backend": "pyav",
105
+ "streaming": false
106
+ },
107
+ "env": null,
108
+ "policy": {
109
+ "type": "act",
110
+ "n_obs_steps": 1,
111
+ "input_features": {
112
+ "observation.state": {
113
+ "type": "STATE",
114
+ "shape": [
115
+ 14
116
+ ]
117
+ },
118
+ "observation.images.Cam1_global": {
119
+ "type": "VISUAL",
120
+ "shape": [
121
+ 3,
122
+ 256,
123
+ 256
124
+ ]
125
+ },
126
+ "observation.images.Cam1_local": {
127
+ "type": "VISUAL",
128
+ "shape": [
129
+ 3,
130
+ 256,
131
+ 256
132
+ ]
133
+ },
134
+ "observation.images.Cam1_red": {
135
+ "type": "VISUAL",
136
+ "shape": [
137
+ 3,
138
+ 256,
139
+ 256
140
+ ]
141
+ },
142
+ "observation.images.Cam2_global": {
143
+ "type": "VISUAL",
144
+ "shape": [
145
+ 3,
146
+ 256,
147
+ 256
148
+ ]
149
+ },
150
+ "observation.images.Cam2_local": {
151
+ "type": "VISUAL",
152
+ "shape": [
153
+ 3,
154
+ 256,
155
+ 256
156
+ ]
157
+ },
158
+ "observation.images.Cam2_red": {
159
+ "type": "VISUAL",
160
+ "shape": [
161
+ 3,
162
+ 256,
163
+ 256
164
+ ]
165
+ }
166
+ },
167
+ "output_features": {
168
+ "action": {
169
+ "type": "ACTION",
170
+ "shape": [
171
+ 6
172
+ ]
173
+ }
174
+ },
175
+ "device": "cuda",
176
+ "use_amp": false,
177
+ "use_peft": false,
178
+ "push_to_hub": false,
179
+ "repo_id": "act_pick5_drq",
180
+ "private": null,
181
+ "tags": null,
182
+ "license": null,
183
+ "pretrained_path": null,
184
+ "chunk_size": 100,
185
+ "n_action_steps": 100,
186
+ "normalization_mapping": {
187
+ "VISUAL": "MEAN_STD",
188
+ "STATE": "MEAN_STD",
189
+ "ACTION": "MEAN_STD"
190
+ },
191
+ "vision_backbone": "resnet18",
192
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
193
+ "replace_final_stride_with_dilation": false,
194
+ "pre_norm": false,
195
+ "dim_model": 512,
196
+ "n_heads": 8,
197
+ "dim_feedforward": 3200,
198
+ "feedforward_activation": "relu",
199
+ "n_encoder_layers": 4,
200
+ "n_decoder_layers": 1,
201
+ "use_vae": true,
202
+ "latent_dim": 32,
203
+ "n_vae_encoder_layers": 4,
204
+ "temporal_ensemble_coeff": null,
205
+ "dropout": 0.1,
206
+ "kl_weight": 10.0,
207
+ "optimizer_lr": 1e-05,
208
+ "optimizer_weight_decay": 0.0001,
209
+ "optimizer_lr_backbone": 1e-05
210
+ },
211
+ "output_dir": "outputs/train/2026-02-11/16-37-18_act_pick5_drq",
212
+ "job_name": "act_pick5_drq",
213
+ "resume": false,
214
+ "seed": 1000,
215
+ "num_workers": 12,
216
+ "batch_size": 8,
217
+ "steps": 1000000,
218
+ "eval_freq": 20000,
219
+ "log_freq": 200,
220
+ "tolerance_s": 0.0001,
221
+ "save_checkpoint": true,
222
+ "save_freq": 100000,
223
+ "use_policy_training_preset": true,
224
+ "optimizer": {
225
+ "type": "adamw",
226
+ "lr": 1e-05,
227
+ "weight_decay": 0.0001,
228
+ "grad_clip_norm": 10.0,
229
+ "betas": [
230
+ 0.9,
231
+ 0.999
232
+ ],
233
+ "eps": 1e-08
234
+ },
235
+ "scheduler": null,
236
+ "eval": {
237
+ "n_episodes": 50,
238
+ "batch_size": 50,
239
+ "use_async_envs": false
240
+ },
241
+ "wandb": {
242
+ "enable": true,
243
+ "disable_artifact": false,
244
+ "project": "lerobot_leo",
245
+ "entity": null,
246
+ "notes": null,
247
+ "run_id": "efad68xu",
248
+ "mode": null
249
+ },
250
+ "peft": null,
251
+ "use_rabc": false,
252
+ "rabc_progress_path": null,
253
+ "rabc_kappa": 0.01,
254
+ "rabc_epsilon": 1e-06,
255
+ "rabc_head_mode": "sparse",
256
+ "rename_map": {},
257
+ "checkpoint_path": null
258
+ }
checkpoints/0400000/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,191 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 1e-05,
4
+ "betas": [
5
+ 0.9,
6
+ 0.999
7
+ ],
8
+ "eps": 1e-08,
9
+ "weight_decay": 0.0001,
10
+ "amsgrad": false,
11
+ "maximize": false,
12
+ "foreach": null,
13
+ "capturable": false,
14
+ "differentiable": false,
15
+ "fused": null,
16
+ "decoupled_weight_decay": true,
17
+ "params": [
18
+ 0,
19
+ 1,
20
+ 2,
21
+ 3,
22
+ 4,
23
+ 5,
24
+ 6,
25
+ 7,
26
+ 8,
27
+ 9,
28
+ 10,
29
+ 11,
30
+ 12,
31
+ 13,
32
+ 14,
33
+ 15,
34
+ 16,
35
+ 17,
36
+ 18,
37
+ 19,
38
+ 20,
39
+ 21,
40
+ 22,
41
+ 23,
42
+ 24,
43
+ 25,
44
+ 26,
45
+ 27,
46
+ 28,
47
+ 29,
48
+ 30,
49
+ 31,
50
+ 32,
51
+ 33,
52
+ 34,
53
+ 35,
54
+ 36,
55
+ 37,
56
+ 38,
57
+ 39,
58
+ 40,
59
+ 41,
60
+ 42,
61
+ 43,
62
+ 44,
63
+ 45,
64
+ 46,
65
+ 47,
66
+ 48,
67
+ 49,
68
+ 50,
69
+ 51,
70
+ 52,
71
+ 53,
72
+ 54,
73
+ 55,
74
+ 56,
75
+ 57,
76
+ 58,
77
+ 59,
78
+ 60,
79
+ 61,
80
+ 62,
81
+ 63,
82
+ 64,
83
+ 65,
84
+ 66,
85
+ 67,
86
+ 68,
87
+ 69,
88
+ 70,
89
+ 71,
90
+ 72,
91
+ 73,
92
+ 74,
93
+ 75,
94
+ 76,
95
+ 77,
96
+ 78,
97
+ 79,
98
+ 80,
99
+ 81,
100
+ 82,
101
+ 83,
102
+ 84,
103
+ 85,
104
+ 86,
105
+ 87,
106
+ 88,
107
+ 89,
108
+ 90,
109
+ 91,
110
+ 92,
111
+ 93,
112
+ 94,
113
+ 95,
114
+ 96,
115
+ 97,
116
+ 98,
117
+ 99,
118
+ 100,
119
+ 101,
120
+ 102,
121
+ 103,
122
+ 104,
123
+ 105,
124
+ 106,
125
+ 107,
126
+ 108,
127
+ 109,
128
+ 110,
129
+ 111,
130
+ 112,
131
+ 113,
132
+ 114,
133
+ 115,
134
+ 116,
135
+ 117,
136
+ 118,
137
+ 119,
138
+ 120,
139
+ 121,
140
+ 122,
141
+ 123,
142
+ 124,
143
+ 125,
144
+ 126,
145
+ 127,
146
+ 128,
147
+ 129,
148
+ 130,
149
+ 131,
150
+ 132
151
+ ]
152
+ },
153
+ {
154
+ "lr": 1e-05,
155
+ "betas": [
156
+ 0.9,
157
+ 0.999
158
+ ],
159
+ "eps": 1e-08,
160
+ "weight_decay": 0.0001,
161
+ "amsgrad": false,
162
+ "maximize": false,
163
+ "foreach": null,
164
+ "capturable": false,
165
+ "differentiable": false,
166
+ "fused": null,
167
+ "decoupled_weight_decay": true,
168
+ "params": [
169
+ 133,
170
+ 134,
171
+ 135,
172
+ 136,
173
+ 137,
174
+ 138,
175
+ 139,
176
+ 140,
177
+ 141,
178
+ 142,
179
+ 143,
180
+ 144,
181
+ 145,
182
+ 146,
183
+ 147,
184
+ 148,
185
+ 149,
186
+ 150,
187
+ 151,
188
+ 152
189
+ ]
190
+ }
191
+ ]
checkpoints/0400000/training_state/optimizer_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c25ac448a096d6bc38dc4808e44a4eb32eb9a6493101a7a96b82fc5dff510558
3
+ size 412883196
checkpoints/0400000/training_state/rng_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b18fae5a71a3135d3fcec7b127e4c37ac90ec62c0e2eab95efb14de137c852d4
3
+ size 15708
checkpoints/0400000/training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 400000
3
+ }
wandb/debug-internal.log ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {"time":"2026-02-11T16:37:22.666598641+11:00","level":"INFO","msg":"stream: starting","core version":"0.24.0"}
2
+ {"time":"2026-02-11T16:37:23.074495728+11:00","level":"INFO","msg":"stream: created new stream","id":"efad68xu"}
3
+ {"time":"2026-02-11T16:37:23.074599598+11:00","level":"INFO","msg":"handler: started","stream_id":"efad68xu"}
4
+ {"time":"2026-02-11T16:37:23.076431641+11:00","level":"INFO","msg":"stream: started","id":"efad68xu"}
5
+ {"time":"2026-02-11T16:37:23.076475937+11:00","level":"INFO","msg":"writer: started","stream_id":"efad68xu"}
6
+ {"time":"2026-02-11T16:37:23.076477822+11:00","level":"INFO","msg":"sender: started","stream_id":"efad68xu"}
wandb/debug.log ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-02-11 16:37:21,698 INFO MainThread:1343881 [wandb_setup.py:_flush():81] Current SDK version is 0.24.0
2
+ 2026-02-11 16:37:21,698 INFO MainThread:1343881 [wandb_setup.py:_flush():81] Configure stats pid to 1343881
3
+ 2026-02-11 16:37:21,698 INFO MainThread:1343881 [wandb_setup.py:_flush():81] Loading settings from environment variables
4
+ 2026-02-11 16:37:21,698 INFO MainThread:1343881 [wandb_init.py:setup_run_log_directory():717] Logging user logs to outputs/train/2026-02-11/16-37-18_act_pick5_drq/wandb/run-20260211_163721-efad68xu/logs/debug.log
5
+ 2026-02-11 16:37:21,698 INFO MainThread:1343881 [wandb_init.py:setup_run_log_directory():718] Logging internal logs to outputs/train/2026-02-11/16-37-18_act_pick5_drq/wandb/run-20260211_163721-efad68xu/logs/debug-internal.log
6
+ 2026-02-11 16:37:21,698 INFO MainThread:1343881 [wandb_init.py:init():844] calling init triggers
7
+ 2026-02-11 16:37:21,698 INFO MainThread:1343881 [wandb_init.py:init():849] wandb.init called with sweep_config: {}
8
+ config: {'dataset': {'repo_id': 'Sendera/pick5', 'root': '/srv/scratch/z5649701/datasets/Sendera/pick5', 'episodes': None, 'image_transforms': {'enable': True, 'max_num_transforms': 3, 'random_order': False, 'tfs': {'brightness': {'weight': 0.0, 'type': 'ColorJitter', 'kwargs': {'brightness': [0.8, 1.2]}}, 'contrast': {'weight': 0.0, 'type': 'ColorJitter', 'kwargs': {'contrast': [0.8, 1.2]}}, 'saturation': {'weight': 0.0, 'type': 'ColorJitter', 'kwargs': {'saturation': [0.5, 1.5]}}, 'hue': {'weight': 0.0, 'type': 'ColorJitter', 'kwargs': {'hue': [-0.05, 0.05]}}, 'sharpness': {'weight': 0.0, 'type': 'SharpnessJitter', 'kwargs': {'sharpness': [0.5, 1.5]}}, 'affine': {'weight': 0.0, 'type': 'RandomAffine', 'kwargs': {'degrees': [-5.0, 5.0], 'translate': [0.05, 0.05]}}, 'resize': {'weight': 0.0, 'type': 'Resize', 'kwargs': {'size': 512, 'interpolation': 2, 'antialias': True}}, 'center_crop': {'weight': 0.0, 'type': 'CenterCrop', 'kwargs': {'size': 512}}, 'random_crop': {'weight': 1.0, 'type': 'RandomCrop', 'kwargs': {'size': 256, 'padding': 13, 'padding_mode': 'edge'}}}}, 'revision': None, 'use_imagenet_stats': True, 'video_backend': 'pyav', 'streaming': False}, 'env': None, 'policy': {'type': 'act', 'n_obs_steps': 1, 'input_features': {}, 'output_features': {}, 'device': 'cuda', 'use_amp': False, 'use_peft': False, 'push_to_hub': False, 'repo_id': 'act_pick5_drq', 'private': None, 'tags': None, 'license': None, 'pretrained_path': None, 'chunk_size': 100, 'n_action_steps': 100, 'normalization_mapping': {'VISUAL': <NormalizationMode.MEAN_STD: 'MEAN_STD'>, 'STATE': <NormalizationMode.MEAN_STD: 'MEAN_STD'>, 'ACTION': <NormalizationMode.MEAN_STD: 'MEAN_STD'>}, 'vision_backbone': 'resnet18', 'pretrained_backbone_weights': 'ResNet18_Weights.IMAGENET1K_V1', 'replace_final_stride_with_dilation': False, 'pre_norm': False, 'dim_model': 512, 'n_heads': 8, 'dim_feedforward': 3200, 'feedforward_activation': 'relu', 'n_encoder_layers': 4, 'n_decoder_layers': 1, 'use_vae': True, 'latent_dim': 32, 'n_vae_encoder_layers': 4, 'temporal_ensemble_coeff': None, 'dropout': 0.1, 'kl_weight': 10.0, 'optimizer_lr': 1e-05, 'optimizer_weight_decay': 0.0001, 'optimizer_lr_backbone': 1e-05}, 'output_dir': 'outputs/train/2026-02-11/16-37-18_act_pick5_drq', 'job_name': 'act_pick5_drq', 'resume': False, 'seed': 1000, 'num_workers': 12, 'batch_size': 8, 'steps': 1000000, 'eval_freq': 20000, 'log_freq': 200, 'tolerance_s': 0.0001, 'save_checkpoint': True, 'save_freq': 100000, 'use_policy_training_preset': True, 'optimizer': {'type': 'adamw', 'lr': 1e-05, 'weight_decay': 0.0001, 'grad_clip_norm': 10.0, 'betas': [0.9, 0.999], 'eps': 1e-08}, 'scheduler': None, 'eval': {'n_episodes': 50, 'batch_size': 50, 'use_async_envs': False}, 'wandb': {'enable': True, 'disable_artifact': False, 'project': 'lerobot_leo', 'entity': None, 'notes': None, 'run_id': None, 'mode': None}, 'peft': None, 'use_rabc': False, 'rabc_progress_path': None, 'rabc_kappa': 0.01, 'rabc_epsilon': 1e-06, 'rabc_head_mode': 'sparse', 'rename_map': {}, 'checkpoint_path': None, '_wandb': {}}
9
+ 2026-02-11 16:37:21,699 INFO MainThread:1343881 [wandb_init.py:init():892] starting backend
10
+ 2026-02-11 16:37:22,648 INFO MainThread:1343881 [wandb_init.py:init():895] sending inform_init request
11
+ 2026-02-11 16:37:22,664 INFO MainThread:1343881 [wandb_init.py:init():903] backend started and connected
12
+ 2026-02-11 16:37:22,665 INFO MainThread:1343881 [wandb_init.py:init():973] updated telemetry
13
+ 2026-02-11 16:37:22,666 INFO MainThread:1343881 [wandb_init.py:init():997] communicating run to backend with 90.0 second timeout
14
+ 2026-02-11 16:37:23,452 INFO MainThread:1343881 [wandb_init.py:init():1044] starting run threads in backend
15
+ 2026-02-11 16:37:23,710 INFO MainThread:1343881 [wandb_run.py:_console_start():2529] atexit reg
16
+ 2026-02-11 16:37:23,711 INFO MainThread:1343881 [wandb_run.py:_redirect():2377] redirect: wrap_raw
17
+ 2026-02-11 16:37:23,711 INFO MainThread:1343881 [wandb_run.py:_redirect():2446] Wrapping output streams.
18
+ 2026-02-11 16:37:23,711 INFO MainThread:1343881 [wandb_run.py:_redirect():2469] Redirects installed.
19
+ 2026-02-11 16:37:23,731 INFO MainThread:1343881 [wandb_init.py:init():1084] run started, returning control to user process
wandb/run-20260211_163721-efad68xu/files/output.log ADDED
The diff for this file is too large to render. See raw diff
 
wandb/run-20260211_163721-efad68xu/files/requirements.txt ADDED
@@ -0,0 +1,123 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ tqdm==4.67.3
2
+ nvidia-cuda-runtime-cu12==12.8.90
3
+ typer-slim==0.21.1
4
+ pandas==2.3.3
5
+ sympy==1.14.0
6
+ httpcore==1.0.9
7
+ pynput==1.8.1
8
+ pfzy==0.3.4
9
+ Jinja2==3.1.6
10
+ exceptiongroup==1.3.1
11
+ filelock==3.20.3
12
+ gitdb==4.0.12
13
+ aiosignal==1.4.0
14
+ nvidia-cufft-cu12==11.3.3.83
15
+ frozenlist==1.8.0
16
+ jsonlines==4.0.0
17
+ multiprocess==0.70.16
18
+ cuda-pathfinder==1.2.2
19
+ inquirerpy==0.3.4
20
+ nvidia-cublas-cu12==12.8.4.1
21
+ torchaudio==2.11.0.dev20260123+cu128
22
+ nvidia-nccl-cu12==2.28.9
23
+ protobuf==6.33.4
24
+ importlib_metadata==8.7.1
25
+ aiohttp==3.13.3
26
+ GitPython==3.1.46
27
+ async-timeout==5.0.1
28
+ annotated-types==0.7.0
29
+ typing-inspect==0.9.0
30
+ nvidia-nvjitlink-cu12==12.8.93
31
+ charset-normalizer==3.4.4
32
+ pyarrow==23.0.0
33
+ six==1.17.0
34
+ nvidia-curand-cu12==10.3.9.90
35
+ cmake==4.1.3
36
+ h11==0.16.0
37
+ zipp==3.23.0
38
+ typing-inspection==0.4.2
39
+ wcwidth==0.3.1
40
+ propcache==0.4.1
41
+ python-xlib==0.33
42
+ multidict==6.7.0
43
+ psutil==7.2.1
44
+ typing_extensions==4.15.0
45
+ nvidia-cusparse-cu12==12.5.8.93
46
+ triton==3.6.0+git9844da95
47
+ datasets==4.1.1
48
+ pydantic_core==2.41.5
49
+ attrs==25.4.0
50
+ yarl==1.22.0
51
+ networkx==3.4.2
52
+ torch==2.11.0.dev20260122+cu128
53
+ evdev==1.9.2
54
+ transformers==4.45.2
55
+ nvidia-cusolver-cu12==11.7.3.90
56
+ tokenizers==0.20.3
57
+ setuptools==80.10.1
58
+ opencv-python-headless==4.12.0.88
59
+ safetensors==0.7.0
60
+ requests==2.32.5
61
+ av==15.1.0
62
+ mypy_extensions==1.1.0
63
+ MarkupSafe==3.0.2
64
+ imageio-ffmpeg==0.6.0
65
+ pydantic==2.12.5
66
+ sentry-sdk==2.50.0
67
+ pip==25.3
68
+ rerun-sdk==0.26.2
69
+ pillow==12.1.0
70
+ gymnasium==1.2.3
71
+ termcolor==3.3.0
72
+ lerobot==0.4.4
73
+ wandb==0.24.0
74
+ torchvision==0.25.0.dev20260122+cu128
75
+ xxhash==3.6.0
76
+ PyYAML==6.0.3
77
+ accelerate==1.12.0
78
+ deepdiff==8.6.1
79
+ nvidia-cuda-cupti-cu12==12.8.90
80
+ idna==3.11
81
+ prompt_toolkit==3.0.52
82
+ numpy==2.2.6
83
+ mpmath==1.3.0
84
+ einops==0.8.1
85
+ urllib3==2.6.3
86
+ hf-xet==1.2.0
87
+ nvidia-nvtx-cu12==12.8.90
88
+ fsspec==2026.1.0
89
+ ImageIO==2.37.2
90
+ tzdata==2025.3
91
+ click==8.3.1
92
+ nvidia-cudnn-cu12==9.15.1.9
93
+ nvidia-cufile-cu12==1.13.1.3
94
+ diffusers==0.35.2
95
+ cloudpickle==3.1.2
96
+ pyyaml-include==1.4.1
97
+ packaging==26.0
98
+ Farama-Notifications==0.0.4
99
+ torchcodec==0.10.0
100
+ pyserial==3.5
101
+ certifi==2026.1.4
102
+ smmap==5.0.2
103
+ draccus==0.10.0
104
+ mergedeep==1.3.4
105
+ nvidia-cuda-nvrtc-cu12==12.8.93
106
+ regex==2026.1.15
107
+ pytz==2025.2
108
+ aiohappyeyeballs==2.6.1
109
+ nvidia-nvshmem-cu12==3.4.5
110
+ python-dateutil==2.9.0.post0
111
+ toml==0.10.2
112
+ platformdirs==4.5.1
113
+ shellingham==1.5.4
114
+ huggingface_hub==0.36.1
115
+ httpx==0.28.1
116
+ hf_transfer==0.1.9
117
+ wheel==0.46.3
118
+ cuda-bindings==12.9.4
119
+ orderly-set==5.5.0
120
+ dill==0.4.0
121
+ nvidia-cusparselt-cu12==0.7.1
122
+ anyio==4.12.1
123
+ lerobot==0.4.4
wandb/run-20260211_163721-efad68xu/files/wandb-metadata.json ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "os": "Linux-4.18.0-553.97.1.el8_10.x86_64-x86_64-with-glibc2.28",
3
+ "python": "CPython 3.10.19",
4
+ "startedAt": "2026-02-11T05:37:21.675736Z",
5
+ "args": [
6
+ "--policy.type",
7
+ "act",
8
+ "--policy.repo_id",
9
+ "act_pick5_drq",
10
+ "--policy.push_to_hub",
11
+ "false",
12
+ "--dataset.repo_id",
13
+ "Sendera/pick5",
14
+ "--dataset.root",
15
+ "/srv/scratch/z5649701/datasets/Sendera/pick5",
16
+ "--dataset.video_backend",
17
+ "pyav",
18
+ "--batch_size",
19
+ "8",
20
+ "--steps",
21
+ "1000000",
22
+ "--save_checkpoint",
23
+ "true",
24
+ "--save_freq",
25
+ "100000",
26
+ "--num_workers",
27
+ "12",
28
+ "--job_name",
29
+ "act_pick5_drq",
30
+ "--wandb.enable",
31
+ "true",
32
+ "--wandb.project",
33
+ "lerobot_leo",
34
+ "--dataset.image_transforms.enable=true"
35
+ ],
36
+ "program": "/srv/scratch/z5649701/projects/lerobot_lib/src/lerobot/scripts/lerobot_train.py",
37
+ "root": "outputs/train/2026-02-11/16-37-18_act_pick5_drq",
38
+ "host": "k205",
39
+ "executable": "/srv/scratch/z5649701/micromamba/envs/lerobot_6000/bin/python",
40
+ "cpu_count": 112,
41
+ "cpu_count_logical": 112,
42
+ "gpu": "NVIDIA H200",
43
+ "gpu_count": 1,
44
+ "disk": {
45
+ "/": {
46
+ "total": "34342961152",
47
+ "used": "15171817472"
48
+ }
49
+ },
50
+ "memory": {
51
+ "total": "2164004765696"
52
+ },
53
+ "gpu_nvidia": [
54
+ {
55
+ "name": "NVIDIA H200",
56
+ "memoryTotal": "150754820096",
57
+ "cudaCores": 16896,
58
+ "architecture": "Hopper",
59
+ "uuid": "GPU-d1b2094a-b591-8cad-63d0-c7d2b44ee3b9"
60
+ }
61
+ ],
62
+ "cudaVersion": "13.1",
63
+ "writerId": "b73qdid1cg3gx833hefzheqe8rwxhzdg"
64
+ }