Hugging Face's logo

sungkyunner
/
focus-token-100k

Safetensors
Model card Files
xet
 Community
sungkyunner commited on
Commit
e949cf3
·
verified ·
1 Parent(s): a1f3a4d

Upload folder using huggingface_hub

Browse files
Files changed (7) hide show
  1. config.json +96 -0
  2. model.safetensors +3 -0
  3. policy_postprocessor.json +32 -0
  4. policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  5. policy_preprocessor.json +79 -0
  6. policy_preprocessor_step_5_normalizer_processor.safetensors +3 -0
  7. train_config.json +250 -0
config.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "smolvla",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.images.image": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 256,
10
+ 256
11
+ ]
12
+ },
13
+ "observation.images.image2": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 256,
18
+ 256
19
+ ]
20
+ },
21
+ "observation.state": {
22
+ "type": "STATE",
23
+ "shape": [
24
+ 8
25
+ ]
26
+ }
27
+ },
28
+ "output_features": {
29
+ "action": {
30
+ "type": "ACTION",
31
+ "shape": [
32
+ 7
33
+ ]
34
+ }
35
+ },
36
+ "device": "cuda",
37
+ "use_amp": false,
38
+ "use_peft": false,
39
+ "push_to_hub": false,
40
+ "repo_id": null,
41
+ "private": null,
42
+ "tags": null,
43
+ "license": null,
44
+ "pretrained_path": null,
45
+ "pretrained_revision": null,
46
+ "chunk_size": 50,
47
+ "n_action_steps": 50,
48
+ "normalization_mapping": {
49
+ "VISUAL": "IDENTITY",
50
+ "STATE": "MEAN_STD",
51
+ "ACTION": "MEAN_STD"
52
+ },
53
+ "max_state_dim": 32,
54
+ "max_action_dim": 32,
55
+ "resize_imgs_with_padding": [
56
+ 512,
57
+ 512
58
+ ],
59
+ "empty_cameras": 0,
60
+ "adapt_to_pi_aloha": false,
61
+ "use_delta_joint_actions_aloha": false,
62
+ "tokenizer_max_length": 48,
63
+ "num_steps": 10,
64
+ "use_cache": true,
65
+ "freeze_vision_encoder": true,
66
+ "train_expert_only": true,
67
+ "train_state_proj": true,
68
+ "optimizer_lr": 0.0001,
69
+ "optimizer_betas": [
70
+ 0.9,
71
+ 0.95
72
+ ],
73
+ "optimizer_eps": 1e-08,
74
+ "optimizer_weight_decay": 1e-10,
75
+ "optimizer_grad_clip_norm": 10.0,
76
+ "scheduler_warmup_steps": 1000,
77
+ "scheduler_decay_steps": 100000,
78
+ "scheduler_decay_lr": 2.5e-06,
79
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
80
+ "load_vlm_weights": true,
81
+ "add_image_special_tokens": false,
82
+ "attention_mode": "cross_attn",
83
+ "prefix_length": -1,
84
+ "pad_language_to": "longest",
85
+ "num_expert_layers": -1,
86
+ "num_vlm_layers": 16,
87
+ "self_attn_every_n_layers": 2,
88
+ "expert_width_multiplier": 0.75,
89
+ "focus_token_keep_ratio": 0.5,
90
+ "focus_token_start_layer": 8,
91
+ "min_period": 0.004,
92
+ "max_period": 4.0,
93
+ "rtc_config": null,
94
+ "compile_model": false,
95
+ "compile_mode": "max-autotune"
96
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f281106919fe969b1d1fd9ac4ded25e65948725ef2745d544cedcaac8023a7f1
3
+ size 906712520
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
+ }
policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a002c0df7f79c5b169c5a899ad151d4ea1bed246c7d82bd93ed1556558d517a9
3
+ size 7668
policy_preprocessor.json ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "longest",
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.image": {
42
+ "type": "VISUAL",
43
+ "shape": [
44
+ 3,
45
+ 256,
46
+ 256
47
+ ]
48
+ },
49
+ "observation.images.image2": {
50
+ "type": "VISUAL",
51
+ "shape": [
52
+ 3,
53
+ 256,
54
+ 256
55
+ ]
56
+ },
57
+ "observation.state": {
58
+ "type": "STATE",
59
+ "shape": [
60
+ 8
61
+ ]
62
+ },
63
+ "action": {
64
+ "type": "ACTION",
65
+ "shape": [
66
+ 7
67
+ ]
68
+ }
69
+ },
70
+ "norm_map": {
71
+ "VISUAL": "IDENTITY",
72
+ "STATE": "MEAN_STD",
73
+ "ACTION": "MEAN_STD"
74
+ }
75
+ },
76
+ "state_file": "policy_preprocessor_step_5_normalizer_processor.safetensors"
77
+ }
78
+ ]
79
+ }
policy_preprocessor_step_5_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f98234b21ce8ffacf1b86657f3cf8e24ac6c31ebcd6a3b83cecd484439674306
3
+ size 7676
train_config.json ADDED
@@ -0,0 +1,250 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "HuggingFaceVLA/libero",
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": "86958911c0f959db2bbbdb107eb3e17c5f9c798e",
78
+ "use_imagenet_stats": true,
79
+ "video_backend": "torchcodec",
80
+ "return_uint8": false,
81
+ "depth_output_unit": "mm",
82
+ "streaming": false,
83
+ "eval_split": 0.0
84
+ },
85
+ "env": null,
86
+ "policy": {
87
+ "type": "smolvla",
88
+ "n_obs_steps": 1,
89
+ "input_features": {
90
+ "observation.images.image": {
91
+ "type": "VISUAL",
92
+ "shape": [
93
+ 3,
94
+ 256,
95
+ 256
96
+ ]
97
+ },
98
+ "observation.images.image2": {
99
+ "type": "VISUAL",
100
+ "shape": [
101
+ 3,
102
+ 256,
103
+ 256
104
+ ]
105
+ },
106
+ "observation.state": {
107
+ "type": "STATE",
108
+ "shape": [
109
+ 8
110
+ ]
111
+ }
112
+ },
113
+ "output_features": {
114
+ "action": {
115
+ "type": "ACTION",
116
+ "shape": [
117
+ 7
118
+ ]
119
+ }
120
+ },
121
+ "device": "cuda",
122
+ "use_amp": false,
123
+ "use_peft": false,
124
+ "push_to_hub": false,
125
+ "repo_id": null,
126
+ "private": null,
127
+ "tags": null,
128
+ "license": null,
129
+ "pretrained_path": null,
130
+ "pretrained_revision": null,
131
+ "chunk_size": 50,
132
+ "n_action_steps": 50,
133
+ "normalization_mapping": {
134
+ "VISUAL": "IDENTITY",
135
+ "STATE": "MEAN_STD",
136
+ "ACTION": "MEAN_STD"
137
+ },
138
+ "max_state_dim": 32,
139
+ "max_action_dim": 32,
140
+ "resize_imgs_with_padding": [
141
+ 512,
142
+ 512
143
+ ],
144
+ "empty_cameras": 0,
145
+ "adapt_to_pi_aloha": false,
146
+ "use_delta_joint_actions_aloha": false,
147
+ "tokenizer_max_length": 48,
148
+ "num_steps": 10,
149
+ "use_cache": true,
150
+ "freeze_vision_encoder": true,
151
+ "train_expert_only": true,
152
+ "train_state_proj": true,
153
+ "optimizer_lr": 0.0001,
154
+ "optimizer_betas": [
155
+ 0.9,
156
+ 0.95
157
+ ],
158
+ "optimizer_eps": 1e-08,
159
+ "optimizer_weight_decay": 1e-10,
160
+ "optimizer_grad_clip_norm": 10.0,
161
+ "scheduler_warmup_steps": 1000,
162
+ "scheduler_decay_steps": 100000,
163
+ "scheduler_decay_lr": 2.5e-06,
164
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
165
+ "load_vlm_weights": true,
166
+ "add_image_special_tokens": false,
167
+ "attention_mode": "cross_attn",
168
+ "prefix_length": -1,
169
+ "pad_language_to": "longest",
170
+ "num_expert_layers": -1,
171
+ "num_vlm_layers": 16,
172
+ "self_attn_every_n_layers": 2,
173
+ "expert_width_multiplier": 0.75,
174
+ "focus_token_keep_ratio": 0.5,
175
+ "focus_token_start_layer": 8,
176
+ "min_period": 0.004,
177
+ "max_period": 4.0,
178
+ "rtc_config": null,
179
+ "compile_model": false,
180
+ "compile_mode": "max-autotune"
181
+ },
182
+ "reward_model": null,
183
+ "output_dir": "experiments/focus_token/results/20260804T133556Z_focus50_seed_1000_100k/focus50/seed_1000/train",
184
+ "job_name": "focus_token_focus50_1000_100k",
185
+ "resume": false,
186
+ "seed": 1000,
187
+ "cudnn_deterministic": true,
188
+ "num_workers": 4,
189
+ "batch_size": 4,
190
+ "prefetch_factor": 4,
191
+ "persistent_workers": true,
192
+ "dataloader_multiprocessing_context": "spawn",
193
+ "steps": 100000,
194
+ "env_eval_freq": 0,
195
+ "log_freq": 100,
196
+ "eval_steps": 0,
197
+ "max_eval_samples": 0,
198
+ "tolerance_s": 0.0001,
199
+ "save_checkpoint": true,
200
+ "save_freq": 10000,
201
+ "use_policy_training_preset": true,
202
+ "optimizer": {
203
+ "type": "adamw",
204
+ "lr": 0.0001,
205
+ "weight_decay": 1e-10,
206
+ "grad_clip_norm": 10.0,
207
+ "betas": [
208
+ 0.9,
209
+ 0.95
210
+ ],
211
+ "eps": 1e-08
212
+ },
213
+ "scheduler": {
214
+ "type": "cosine_decay_with_warmup",
215
+ "num_warmup_steps": 1000,
216
+ "num_decay_steps": 100000,
217
+ "peak_lr": 0.0001,
218
+ "decay_lr": 2.5e-06
219
+ },
220
+ "eval": {
221
+ "n_episodes": 50,
222
+ "batch_size": 19,
223
+ "use_async_envs": true,
224
+ "recording": false,
225
+ "recording_repo_id": null,
226
+ "recording_private": false
227
+ },
228
+ "wandb": {
229
+ "enable": false,
230
+ "disable_artifact": false,
231
+ "project": "lerobot",
232
+ "entity": null,
233
+ "notes": null,
234
+ "run_id": null,
235
+ "mode": null,
236
+ "add_tags": true
237
+ },
238
+ "peft": null,
239
+ "job": {
240
+ "target": null,
241
+ "image": "huggingface/lerobot-gpu:latest",
242
+ "timeout": "2d",
243
+ "detach": false,
244
+ "tags": []
245
+ },
246
+ "save_checkpoint_to_hub": false,
247
+ "sample_weighting": null,
248
+ "rename_map": {},
249
+ "checkpoint_path": null
250
+ }