Hugging Face's logo

jasontchan
/
emg-task3-edit

Safetensors
Model card Files
xet
Community
jasontchan commited on
Commit
97c72af
·
verified ·
1 Parent(s): 136517a

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. config.json +110 -0
  2. model.safetensors +3 -0
  3. train_config.json +223 -0
config.json ADDED
@@ -0,0 +1,110 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "smolvla",
3
+ "n_obs_steps": 1,
4
+ "normalization_mapping": {
5
+ "VISUAL": "IDENTITY",
6
+ "STATE": "MEAN_STD",
7
+ "ACTION": "MEAN_STD"
8
+ },
9
+ "input_features": {
10
+ "observation.state": {
11
+ "type": "STATE",
12
+ "shape": [
13
+ 6
14
+ ]
15
+ },
16
+ "observation.images.front": {
17
+ "type": "VISUAL",
18
+ "shape": [
19
+ 3,
20
+ 1080,
21
+ 1920
22
+ ]
23
+ },
24
+ "observation.images.top": {
25
+ "type": "VISUAL",
26
+ "shape": [
27
+ 3,
28
+ 720,
29
+ 1280
30
+ ]
31
+ },
32
+ "observation.images.gripper": {
33
+ "type": "VISUAL",
34
+ "shape": [
35
+ 3,
36
+ 1080,
37
+ 1920
38
+ ]
39
+ },
40
+ "observation.emg.lower": {
41
+ "type": "EMG",
42
+ "shape": [
43
+ 100,
44
+ 8
45
+ ]
46
+ }
47
+ },
48
+ "output_features": {
49
+ "action": {
50
+ "type": "ACTION",
51
+ "shape": [
52
+ 6
53
+ ]
54
+ }
55
+ },
56
+ "device": "cuda",
57
+ "use_amp": false,
58
+ "emg_mask_prob": 0.0,
59
+ "emg_spectrogram": false,
60
+ "n_fft": 64,
61
+ "emg_window_size": 50,
62
+ "emg_n_bins": 10,
63
+ "emg_gate": 0.0,
64
+ "emg_shift": 0,
65
+ "use_camera_keys": {
66
+ "observation.images.top": true,
67
+ "observation.images.gripper": true,
68
+ "observation.images.front": true
69
+ },
70
+ "chunk_size": 50,
71
+ "n_action_steps": 50,
72
+ "max_state_dim": 32,
73
+ "max_action_dim": 32,
74
+ "resize_imgs_with_padding": [
75
+ 512,
76
+ 512
77
+ ],
78
+ "empty_cameras": 0,
79
+ "adapt_to_pi_aloha": false,
80
+ "use_delta_joint_actions_aloha": false,
81
+ "tokenizer_max_length": 48,
82
+ "num_steps": 10,
83
+ "use_cache": true,
84
+ "freeze_vision_encoder": true,
85
+ "train_expert_only": true,
86
+ "train_state_proj": true,
87
+ "optimizer_lr": 0.0001,
88
+ "optimizer_betas": [
89
+ 0.9,
90
+ 0.95
91
+ ],
92
+ "optimizer_eps": 1e-08,
93
+ "optimizer_weight_decay": 1e-10,
94
+ "optimizer_grad_clip_norm": 10.0,
95
+ "scheduler_warmup_steps": 1000,
96
+ "scheduler_decay_steps": 30000,
97
+ "scheduler_decay_lr": 2.5e-06,
98
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
99
+ "load_vlm_weights": true,
100
+ "add_image_special_tokens": false,
101
+ "attention_mode": "cross_attn",
102
+ "prefix_length": 0,
103
+ "pad_language_to": "max_length",
104
+ "num_expert_layers": 0,
105
+ "num_vlm_layers": 16,
106
+ "self_attn_every_n_layers": 2,
107
+ "expert_width_multiplier": 0.75,
108
+ "min_period": 0.004,
109
+ "max_period": 4.0
110
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d35598478bb1866c148bfc31fbbc2c96187a7a69211ff96daac2130781ad1055
3
+ size 907243332
train_config.json ADDED
@@ -0,0 +1,223 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "jasontchan/emg-gestures-task3-edit",
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
+ }
62
+ },
63
+ "revision": null,
64
+ "use_imagenet_stats": true,
65
+ "video_backend": "torchcodec"
66
+ },
67
+ "env": null,
68
+ "policy": {
69
+ "type": "smolvla",
70
+ "n_obs_steps": 1,
71
+ "normalization_mapping": {
72
+ "VISUAL": "IDENTITY",
73
+ "STATE": "MEAN_STD",
74
+ "ACTION": "MEAN_STD"
75
+ },
76
+ "input_features": {
77
+ "observation.state": {
78
+ "type": "STATE",
79
+ "shape": [
80
+ 6
81
+ ]
82
+ },
83
+ "observation.images.front": {
84
+ "type": "VISUAL",
85
+ "shape": [
86
+ 3,
87
+ 1080,
88
+ 1920
89
+ ]
90
+ },
91
+ "observation.images.top": {
92
+ "type": "VISUAL",
93
+ "shape": [
94
+ 3,
95
+ 720,
96
+ 1280
97
+ ]
98
+ },
99
+ "observation.images.gripper": {
100
+ "type": "VISUAL",
101
+ "shape": [
102
+ 3,
103
+ 1080,
104
+ 1920
105
+ ]
106
+ },
107
+ "observation.emg.lower": {
108
+ "type": "EMG",
109
+ "shape": [
110
+ 100,
111
+ 8
112
+ ]
113
+ }
114
+ },
115
+ "output_features": {
116
+ "action": {
117
+ "type": "ACTION",
118
+ "shape": [
119
+ 6
120
+ ]
121
+ }
122
+ },
123
+ "device": "cuda",
124
+ "use_amp": false,
125
+ "emg_mask_prob": 0.0,
126
+ "emg_spectrogram": false,
127
+ "n_fft": 64,
128
+ "emg_window_size": 50,
129
+ "emg_n_bins": 10,
130
+ "emg_gate": 0.0,
131
+ "emg_shift": 0,
132
+ "use_camera_keys": {
133
+ "observation.images.top": true,
134
+ "observation.images.gripper": true,
135
+ "observation.images.front": true
136
+ },
137
+ "chunk_size": 50,
138
+ "n_action_steps": 50,
139
+ "max_state_dim": 32,
140
+ "max_action_dim": 32,
141
+ "resize_imgs_with_padding": [
142
+ 512,
143
+ 512
144
+ ],
145
+ "empty_cameras": 0,
146
+ "adapt_to_pi_aloha": false,
147
+ "use_delta_joint_actions_aloha": false,
148
+ "tokenizer_max_length": 48,
149
+ "num_steps": 10,
150
+ "use_cache": true,
151
+ "freeze_vision_encoder": true,
152
+ "train_expert_only": true,
153
+ "train_state_proj": true,
154
+ "optimizer_lr": 0.0001,
155
+ "optimizer_betas": [
156
+ 0.9,
157
+ 0.95
158
+ ],
159
+ "optimizer_eps": 1e-08,
160
+ "optimizer_weight_decay": 1e-10,
161
+ "optimizer_grad_clip_norm": 10.0,
162
+ "scheduler_warmup_steps": 1000,
163
+ "scheduler_decay_steps": 30000,
164
+ "scheduler_decay_lr": 2.5e-06,
165
+ "vlm_model_name": "HuggingFaceTB/SmolVLM2-500M-Video-Instruct",
166
+ "load_vlm_weights": true,
167
+ "add_image_special_tokens": false,
168
+ "attention_mode": "cross_attn",
169
+ "prefix_length": 0,
170
+ "pad_language_to": "max_length",
171
+ "num_expert_layers": 0,
172
+ "num_vlm_layers": 16,
173
+ "self_attn_every_n_layers": 2,
174
+ "expert_width_multiplier": 0.75,
175
+ "min_period": 0.004,
176
+ "max_period": 4.0
177
+ },
178
+ "output_dir": "outputs/train/emg-task3-edit",
179
+ "job_name": "emg-task3-edit",
180
+ "resume": false,
181
+ "seed": 1000,
182
+ "num_workers": 4,
183
+ "batch_size": 8,
184
+ "steps": 100000,
185
+ "eval_freq": 20000,
186
+ "log_freq": 200,
187
+ "save_checkpoint": true,
188
+ "save_freq": 20000,
189
+ "use_policy_training_preset": true,
190
+ "optimizer": {
191
+ "type": "adamw",
192
+ "lr": 0.0001,
193
+ "weight_decay": 1e-10,
194
+ "grad_clip_norm": 10.0,
195
+ "betas": [
196
+ 0.9,
197
+ 0.95
198
+ ],
199
+ "eps": 1e-08
200
+ },
201
+ "scheduler": {
202
+ "type": "cosine_decay_with_warmup",
203
+ "num_warmup_steps": 1000,
204
+ "num_decay_steps": 30000,
205
+ "peak_lr": 0.0001,
206
+ "decay_lr": 2.5e-06
207
+ },
208
+ "eval": {
209
+ "n_episodes": 50,
210
+ "batch_size": 50,
211
+ "use_async_envs": false
212
+ },
213
+ "wandb": {
214
+ "enable": true,
215
+ "disable_artifact": false,
216
+ "project": "lerobot",
217
+ "entity": null,
218
+ "notes": null,
219
+ "run_id": "kljovo3e",
220
+ "mode": null
221
+ },
222
+ "use_emg": true
223
+ }