Hugging Face's logo

ralfroemer
/
al-hardware-checkpoints

Safetensors
Model card Files
xet
 Community
ralfroemer commited on
Commit
2912665
·
verified ·
1 Parent(s): d6ca0db

real_t3_weighted50_eps10_step10000_s45 round 17 member 1

Browse files
Files changed (7) hide show
  1. real_t3_weighted50_eps10_step10000_s45/round_017/member_01/config.json +123 -0
  2. real_t3_weighted50_eps10_step10000_s45/round_017/member_01/model.safetensors +3 -0
  3. real_t3_weighted50_eps10_step10000_s45/round_017/member_01/policy_postprocessor.json +32 -0
  4. real_t3_weighted50_eps10_step10000_s45/round_017/member_01/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  5. real_t3_weighted50_eps10_step10000_s45/round_017/member_01/policy_preprocessor.json +103 -0
  6. real_t3_weighted50_eps10_step10000_s45/round_017/member_01/policy_preprocessor_step_5_normalizer_processor.safetensors +3 -0
  7. real_t3_weighted50_eps10_step10000_s45/round_017/member_01/train_config.json +222 -0
real_t3_weighted50_eps10_step10000_s45/round_017/member_01/config.json ADDED
@@ -0,0 +1,123 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "smolvla",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.images.primary": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 256,
10
+ 256
11
+ ]
12
+ },
13
+ "observation.images.wrist": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 256,
18
+ 256
19
+ ]
20
+ },
21
+ "observation.state.cartesian": {
22
+ "type": "STATE",
23
+ "shape": [
24
+ 6
25
+ ]
26
+ },
27
+ "observation.state.gripper": {
28
+ "type": "STATE",
29
+ "shape": [
30
+ 1
31
+ ]
32
+ },
33
+ "observation.state.joints": {
34
+ "type": "STATE",
35
+ "shape": [
36
+ 7
37
+ ]
38
+ },
39
+ "observation.state.target": {
40
+ "type": "STATE",
41
+ "shape": [
42
+ 6
43
+ ]
44
+ },
45
+ "observation.state": {
46
+ "type": "STATE",
47
+ "shape": [
48
+ 20
49
+ ]
50
+ }
51
+ },
52
+ "output_features": {
53
+ "action": {
54
+ "type": "ACTION",
55
+ "shape": [
56
+ 7
57
+ ]
58
+ }
59
+ },
60
+ "device": "cuda",
61
+ "use_amp": false,
62
+ "push_to_hub": true,
63
+ "repo_id": null,
64
+ "private": null,
65
+ "tags": null,
66
+ "license": null,
67
+ "pretrained_path": "${STORAGE_ROOT}/outputs/pretrain/hardware_pretrain_v2/seed_4/checkpoints/last/pretrained_model",
68
+ "chunk_size": 50,
69
+ "n_action_steps": 50,
70
+ "normalization_mapping": {
71
+ "VISUAL": "IDENTITY",
72
+ "STATE": "MEAN_STD",
73
+ "ACTION": "MEAN_STD"
74
+ },
75
+ "max_state_dim": 32,
76
+ "max_action_dim": 32,
77
+ "state_feature_indices": [
78
+ 0,
79
+ 1,
80
+ 2,
81
+ 3,
82
+ 4,
83
+ 5,
84
+ 6
85
+ ],
86
+ "resize_imgs_with_padding": [
87
+ 512,
88
+ 512
89
+ ],
90
+ "empty_cameras": 0,
91
+ "adapt_to_pi_aloha": false,
92
+ "use_delta_joint_actions_aloha": false,
93
+ "tokenizer_max_length": 48,
94
+ "num_steps": 10,
95
+ "use_cache": true,
96
+ "freeze_vision_encoder": false,
97
+ "freeze_text_model": false,
98
+ "train_state_proj": true,
99
+ "optimizer_lr": 5e-05,
100
+ "optimizer_betas": [
101
+ 0.9,
102
+ 0.95
103
+ ],
104
+ "optimizer_eps": 1e-08,
105
+ "optimizer_weight_decay": 1e-10,
106
+ "optimizer_grad_clip_norm": 10.0,
107
+ "scheduler_warmup_steps": 500,
108
+ "scheduler_decay_steps": 10000,
109
+ "scheduler_decay_lr": 5e-06,
110
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
111
+ "load_vlm_weights": true,
112
+ "add_image_special_tokens": false,
113
+ "attention_mode": "cross_attn",
114
+ "prefix_length": 0,
115
+ "pad_language_to": "max_length",
116
+ "num_expert_layers": 0,
117
+ "num_vlm_layers": 16,
118
+ "self_attn_every_n_layers": 2,
119
+ "expert_width_multiplier": 0.75,
120
+ "min_period": 0.004,
121
+ "max_period": 4.0,
122
+ "rtc_config": null
123
+ }
real_t3_weighted50_eps10_step10000_s45/round_017/member_01/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a470a3f2095877cded2eb6e5a24fedd328ce989e60488d77b6d837df15614b3
3
+ size 906712520
real_t3_weighted50_eps10_step10000_s45/round_017/member_01/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": "IDENTITY",
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
+ }
real_t3_weighted50_eps10_step10000_s45/round_017/member_01/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5a323cc779fe1f7384982c0970bfac115b90d9467963bbbefa8c0ff4ff99bb5
3
+ size 6060
real_t3_weighted50_eps10_step10000_s45/round_017/member_01/policy_preprocessor.json ADDED
@@ -0,0 +1,103 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "smolvla_new_line_processor",
16
+ "config": {}
17
+ },
18
+ {
19
+ "registry_name": "tokenizer_processor",
20
+ "config": {
21
+ "max_length": 48,
22
+ "task_key": "task",
23
+ "padding_side": "right",
24
+ "padding": "max_length",
25
+ "truncation": true,
26
+ "tokenizer_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct"
27
+ }
28
+ },
29
+ {
30
+ "registry_name": "device_processor",
31
+ "config": {
32
+ "device": "cuda",
33
+ "float_dtype": null
34
+ }
35
+ },
36
+ {
37
+ "registry_name": "normalizer_processor",
38
+ "config": {
39
+ "eps": 1e-08,
40
+ "features": {
41
+ "observation.images.primary": {
42
+ "type": "VISUAL",
43
+ "shape": [
44
+ 3,
45
+ 256,
46
+ 256
47
+ ]
48
+ },
49
+ "observation.images.wrist": {
50
+ "type": "VISUAL",
51
+ "shape": [
52
+ 3,
53
+ 256,
54
+ 256
55
+ ]
56
+ },
57
+ "observation.state.cartesian": {
58
+ "type": "STATE",
59
+ "shape": [
60
+ 6
61
+ ]
62
+ },
63
+ "observation.state.gripper": {
64
+ "type": "STATE",
65
+ "shape": [
66
+ 1
67
+ ]
68
+ },
69
+ "observation.state.joints": {
70
+ "type": "STATE",
71
+ "shape": [
72
+ 7
73
+ ]
74
+ },
75
+ "observation.state.target": {
76
+ "type": "STATE",
77
+ "shape": [
78
+ 6
79
+ ]
80
+ },
81
+ "observation.state": {
82
+ "type": "STATE",
83
+ "shape": [
84
+ 20
85
+ ]
86
+ },
87
+ "action": {
88
+ "type": "ACTION",
89
+ "shape": [
90
+ 7
91
+ ]
92
+ }
93
+ },
94
+ "norm_map": {
95
+ "VISUAL": "IDENTITY",
96
+ "STATE": "MEAN_STD",
97
+ "ACTION": "MEAN_STD"
98
+ }
99
+ },
100
+ "state_file": "policy_preprocessor_step_5_normalizer_processor.safetensors"
101
+ }
102
+ ]
103
+ }
real_t3_weighted50_eps10_step10000_s45/round_017/member_01/policy_preprocessor_step_5_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d5a323cc779fe1f7384982c0970bfac115b90d9467963bbbefa8c0ff4ff99bb5
3
+ size 6060
real_t3_weighted50_eps10_step10000_s45/round_017/member_01/train_config.json ADDED
@@ -0,0 +1,222 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "continuallearning/al_train_pretrainv2_real014",
4
+ "root": null,
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": true,
8
+ "max_num_transforms": 1,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "crop": {
12
+ "weight": 1.0,
13
+ "type": "RandomResizedCrop",
14
+ "kwargs": {
15
+ "size": [
16
+ 256,
17
+ 256
18
+ ],
19
+ "scale": [
20
+ 0.9,
21
+ 1.0
22
+ ],
23
+ "ratio": [
24
+ 0.9,
25
+ 1.1
26
+ ]
27
+ }
28
+ }
29
+ }
30
+ },
31
+ "revision": null,
32
+ "use_imagenet_stats": true,
33
+ "video_backend": "torchcodec",
34
+ "streaming": false,
35
+ "libero_tasks": null,
36
+ "metaworld_tasks": null,
37
+ "pusht_tasks": null,
38
+ "robocasa_tasks": null,
39
+ "real_tasks": {
40
+ "0_put_bowl": "continuallearning/real_0_put_bowl_filtered_consolidated",
41
+ "1_stack_bowls": "continuallearning/real_1_stack_bowls_filtered_consolidated",
42
+ "4_put_lego_into_drawer": "continuallearning/real_4_put_lego_into_drawer_filtered_consolidated"
43
+ },
44
+ "train_repo_id": "continuallearning/al_train_pretrainv2_real014"
45
+ },
46
+ "env": null,
47
+ "policy": {
48
+ "type": "smolvla",
49
+ "n_obs_steps": 1,
50
+ "input_features": {
51
+ "observation.images.primary": {
52
+ "type": "VISUAL",
53
+ "shape": [
54
+ 3,
55
+ 256,
56
+ 256
57
+ ]
58
+ },
59
+ "observation.images.wrist": {
60
+ "type": "VISUAL",
61
+ "shape": [
62
+ 3,
63
+ 256,
64
+ 256
65
+ ]
66
+ },
67
+ "observation.state.cartesian": {
68
+ "type": "STATE",
69
+ "shape": [
70
+ 6
71
+ ]
72
+ },
73
+ "observation.state.gripper": {
74
+ "type": "STATE",
75
+ "shape": [
76
+ 1
77
+ ]
78
+ },
79
+ "observation.state.joints": {
80
+ "type": "STATE",
81
+ "shape": [
82
+ 7
83
+ ]
84
+ },
85
+ "observation.state.target": {
86
+ "type": "STATE",
87
+ "shape": [
88
+ 6
89
+ ]
90
+ },
91
+ "observation.state": {
92
+ "type": "STATE",
93
+ "shape": [
94
+ 20
95
+ ]
96
+ }
97
+ },
98
+ "output_features": {
99
+ "action": {
100
+ "type": "ACTION",
101
+ "shape": [
102
+ 7
103
+ ]
104
+ }
105
+ },
106
+ "device": "cuda",
107
+ "use_amp": false,
108
+ "push_to_hub": true,
109
+ "repo_id": null,
110
+ "private": null,
111
+ "tags": null,
112
+ "license": null,
113
+ "pretrained_path": "/ceph/hpc/data/d2026d06-267-users/euralfr_active_learning/outputs/active_learning/real_robot/real_t3_weighted50_eps10_step10000_s45/round_016/training/member_01/checkpoints/last/pretrained_model",
114
+ "chunk_size": 50,
115
+ "n_action_steps": 50,
116
+ "normalization_mapping": {
117
+ "VISUAL": "IDENTITY",
118
+ "STATE": "MEAN_STD",
119
+ "ACTION": "MEAN_STD"
120
+ },
121
+ "max_state_dim": 32,
122
+ "max_action_dim": 32,
123
+ "state_feature_indices": [
124
+ 0,
125
+ 1,
126
+ 2,
127
+ 3,
128
+ 4,
129
+ 5,
130
+ 6
131
+ ],
132
+ "resize_imgs_with_padding": [
133
+ 512,
134
+ 512
135
+ ],
136
+ "empty_cameras": 0,
137
+ "adapt_to_pi_aloha": false,
138
+ "use_delta_joint_actions_aloha": false,
139
+ "tokenizer_max_length": 48,
140
+ "num_steps": 10,
141
+ "use_cache": true,
142
+ "freeze_vision_encoder": false,
143
+ "freeze_text_model": false,
144
+ "train_state_proj": true,
145
+ "optimizer_lr": 5e-05,
146
+ "optimizer_betas": [
147
+ 0.9,
148
+ 0.95
149
+ ],
150
+ "optimizer_eps": 1e-08,
151
+ "optimizer_weight_decay": 1e-10,
152
+ "optimizer_grad_clip_norm": 10.0,
153
+ "scheduler_warmup_steps": 500,
154
+ "scheduler_decay_steps": 10000,
155
+ "scheduler_decay_lr": 5e-06,
156
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
157
+ "load_vlm_weights": true,
158
+ "add_image_special_tokens": false,
159
+ "attention_mode": "cross_attn",
160
+ "prefix_length": 0,
161
+ "pad_language_to": "max_length",
162
+ "num_expert_layers": 0,
163
+ "num_vlm_layers": 16,
164
+ "self_attn_every_n_layers": 2,
165
+ "expert_width_multiplier": 0.75,
166
+ "min_period": 0.004,
167
+ "max_period": 4.0,
168
+ "rtc_config": null
169
+ },
170
+ "output_dir": "/ceph/hpc/data/d2026d06-267-users/euralfr_active_learning/outputs/active_learning/real_robot/real_t3_weighted50_eps10_step10000_s45/round_017/training/member_01",
171
+ "job_name": "real_t3_weighted50_eps10_step10000_s45_round_017_member_01",
172
+ "resume": false,
173
+ "reinitialize_selected_layers": false,
174
+ "seed": 18003,
175
+ "num_workers": 48,
176
+ "batch_size": 32,
177
+ "steps": 10000,
178
+ "eval_freq": 0,
179
+ "log_freq": 20,
180
+ "save_checkpoint": true,
181
+ "save_freq": 20000,
182
+ "enable_val_loss": true,
183
+ "val_ratio": 0.1,
184
+ "val_freq": null,
185
+ "use_policy_training_preset": true,
186
+ "optimizer": {
187
+ "type": "adamw",
188
+ "lr": 5e-05,
189
+ "weight_decay": 1e-10,
190
+ "grad_clip_norm": 10.0,
191
+ "betas": [
192
+ 0.9,
193
+ 0.95
194
+ ],
195
+ "eps": 1e-08
196
+ },
197
+ "scheduler": {
198
+ "type": "cosine_decay_with_warmup",
199
+ "num_warmup_steps": 500,
200
+ "num_decay_steps": 10000,
201
+ "peak_lr": 5e-05,
202
+ "decay_lr": 5e-06
203
+ },
204
+ "eval": {
205
+ "n_episodes": 1,
206
+ "batch_size": 1,
207
+ "use_async_envs": false
208
+ },
209
+ "wandb": {
210
+ "enable": true,
211
+ "disable_artifact": true,
212
+ "project": "active-vla",
213
+ "entity": "lsy-tum",
214
+ "notes": null,
215
+ "run_id": "y6lfdj93",
216
+ "mode": "online"
217
+ },
218
+ "checkpoint_path": null,
219
+ "rename_map": {},
220
+ "seed_pairs": {},
221
+ "seed_pair": null
222
+ }