Hugging Face's logo

larsvandorp
/
folding_dit

Robotics
LeRobot
Safetensors
so-101
diffusion-policy
multi-task-dit
dinov3
towel-folding
Model card Files
xet
 Community
larsvandorp commited on
Commit
0463b84
·
verified ·
1 Parent(s): 61e34d7

fold6_nocrop_rgray step-28000 (competition DiT, DINOv3-B) at root

Browse files
Files changed (7) hide show
  1. config.json +94 -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 +67 -0
  6. policy_preprocessor_step_4_normalizer_processor.safetensors +3 -0
  7. train_config.json +228 -0
config.json ADDED
@@ -0,0 +1,94 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "multi_task_dit",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 6
9
+ ]
10
+ },
11
+ "observation.images.wrist": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 600,
16
+ 800
17
+ ]
18
+ }
19
+ },
20
+ "output_features": {
21
+ "action": {
22
+ "type": "ACTION",
23
+ "shape": [
24
+ 6
25
+ ]
26
+ }
27
+ },
28
+ "device": "cuda",
29
+ "use_amp": true,
30
+ "use_peft": false,
31
+ "push_to_hub": true,
32
+ "repo_id": "foldingmaster/fold6_nocrop_rgray",
33
+ "private": null,
34
+ "tags": null,
35
+ "license": null,
36
+ "pretrained_path": null,
37
+ "horizon": 32,
38
+ "n_action_steps": 24,
39
+ "objective": "diffusion",
40
+ "noise_scheduler_type": "DDIM",
41
+ "num_train_timesteps": 100,
42
+ "beta_schedule": "squaredcos_cap_v2",
43
+ "beta_start": 0.0001,
44
+ "beta_end": 0.02,
45
+ "prediction_type": "epsilon",
46
+ "clip_sample": true,
47
+ "clip_sample_range": 1.0,
48
+ "num_inference_steps": 10,
49
+ "sigma_min": 0.0,
50
+ "num_integration_steps": 100,
51
+ "integration_method": "euler",
52
+ "timestep_sampling_strategy": "beta",
53
+ "timestep_sampling_s": 0.999,
54
+ "timestep_sampling_alpha": 1.5,
55
+ "timestep_sampling_beta": 1.0,
56
+ "hidden_dim": 512,
57
+ "num_layers": 6,
58
+ "num_heads": 8,
59
+ "dropout": 0.1,
60
+ "use_positional_encoding": false,
61
+ "timestep_embed_dim": 256,
62
+ "use_rope": true,
63
+ "rope_base": 10000.0,
64
+ "vision_encoder_name": "facebook/dinov3-vitb16-pretrain-lvd1689m",
65
+ "use_separate_rgb_encoder_per_camera": false,
66
+ "vision_encoder_lr_multiplier": 0.1,
67
+ "image_resize_shape": [
68
+ 224,
69
+ 304
70
+ ],
71
+ "image_crop_shape": null,
72
+ "image_crop_is_random": true,
73
+ "text_encoder_name": "openai/clip-vit-base-patch16",
74
+ "tokenizer_max_length": 77,
75
+ "tokenizer_padding": "max_length",
76
+ "tokenizer_padding_side": "right",
77
+ "tokenizer_truncation": true,
78
+ "normalization_mapping": {
79
+ "VISUAL": "MEAN_STD",
80
+ "STATE": "MIN_MAX",
81
+ "ACTION": "MIN_MAX"
82
+ },
83
+ "optimizer_lr": 2e-05,
84
+ "optimizer_betas": [
85
+ 0.95,
86
+ 0.999
87
+ ],
88
+ "optimizer_eps": 1e-08,
89
+ "optimizer_weight_decay": 0.0,
90
+ "scheduler_name": "cosine",
91
+ "scheduler_warmup_steps": 0,
92
+ "do_mask_loss_for_padding": false,
93
+ "drop_n_last_frames": 7
94
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22206ea692f517b486021b5e96b36fdeff940609ffe0bf92419d0b1e7b1ff863
3
+ size 881690312
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": "MIN_MAX",
19
+ "ACTION": "MIN_MAX"
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:aa69494a1a3057b93550dd3a84ccc03f5de9e05d90550d31e2878632c5da3029
3
+ size 6560
policy_preprocessor.json ADDED
@@ -0,0 +1,67 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "tokenizer_processor",
16
+ "config": {
17
+ "max_length": 77,
18
+ "task_key": "task",
19
+ "padding_side": "right",
20
+ "padding": "max_length",
21
+ "truncation": true,
22
+ "tokenizer_name": "openai/clip-vit-base-patch16"
23
+ }
24
+ },
25
+ {
26
+ "registry_name": "device_processor",
27
+ "config": {
28
+ "device": "cuda",
29
+ "float_dtype": null
30
+ }
31
+ },
32
+ {
33
+ "registry_name": "normalizer_processor",
34
+ "config": {
35
+ "eps": 1e-08,
36
+ "features": {
37
+ "observation.state": {
38
+ "type": "STATE",
39
+ "shape": [
40
+ 6
41
+ ]
42
+ },
43
+ "observation.images.wrist": {
44
+ "type": "VISUAL",
45
+ "shape": [
46
+ 3,
47
+ 600,
48
+ 800
49
+ ]
50
+ },
51
+ "action": {
52
+ "type": "ACTION",
53
+ "shape": [
54
+ 6
55
+ ]
56
+ }
57
+ },
58
+ "norm_map": {
59
+ "VISUAL": "MEAN_STD",
60
+ "STATE": "MIN_MAX",
61
+ "ACTION": "MIN_MAX"
62
+ }
63
+ },
64
+ "state_file": "policy_preprocessor_step_4_normalizer_processor.safetensors"
65
+ }
66
+ ]
67
+ }
policy_preprocessor_step_4_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:334bb21cd82303c4053843fe695e2b4927c58150697ae6a8ae98e2c9e1a2231d
3
+ size 6568
train_config.json ADDED
@@ -0,0 +1,228 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "foldingmaster/dataset20",
4
+ "root": "./dataset20",
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": true,
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
+ 0.6,
17
+ 1.4
18
+ ]
19
+ }
20
+ },
21
+ "contrast": {
22
+ "weight": 1.0,
23
+ "type": "ColorJitter",
24
+ "kwargs": {
25
+ "contrast": [
26
+ 0.6,
27
+ 1.4
28
+ ]
29
+ }
30
+ },
31
+ "saturation": {
32
+ "weight": 1.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "saturation": [
36
+ 0.0,
37
+ 1.8
38
+ ]
39
+ }
40
+ },
41
+ "hue": {
42
+ "weight": 1.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "hue": [
46
+ -0.25,
47
+ 0.25
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": 0.0,
63
+ "type": "RandomAffine",
64
+ "kwargs": {}
65
+ },
66
+ "random_grayscale": {
67
+ "weight": 1.0,
68
+ "type": "RandomGrayscale",
69
+ "kwargs": {
70
+ "p": 1.0
71
+ }
72
+ }
73
+ }
74
+ },
75
+ "revision": null,
76
+ "use_imagenet_stats": true,
77
+ "video_backend": "pyav",
78
+ "return_uint8": false,
79
+ "streaming": false
80
+ },
81
+ "env": null,
82
+ "policy": {
83
+ "type": "multi_task_dit",
84
+ "n_obs_steps": 2,
85
+ "input_features": {
86
+ "observation.state": {
87
+ "type": "STATE",
88
+ "shape": [
89
+ 6
90
+ ]
91
+ },
92
+ "observation.images.wrist": {
93
+ "type": "VISUAL",
94
+ "shape": [
95
+ 3,
96
+ 600,
97
+ 800
98
+ ]
99
+ }
100
+ },
101
+ "output_features": {
102
+ "action": {
103
+ "type": "ACTION",
104
+ "shape": [
105
+ 6
106
+ ]
107
+ }
108
+ },
109
+ "device": "cuda",
110
+ "use_amp": true,
111
+ "use_peft": false,
112
+ "push_to_hub": true,
113
+ "repo_id": "foldingmaster/fold6_nocrop_rgray",
114
+ "private": null,
115
+ "tags": null,
116
+ "license": null,
117
+ "pretrained_path": null,
118
+ "horizon": 32,
119
+ "n_action_steps": 24,
120
+ "objective": "diffusion",
121
+ "noise_scheduler_type": "DDIM",
122
+ "num_train_timesteps": 100,
123
+ "beta_schedule": "squaredcos_cap_v2",
124
+ "beta_start": 0.0001,
125
+ "beta_end": 0.02,
126
+ "prediction_type": "epsilon",
127
+ "clip_sample": true,
128
+ "clip_sample_range": 1.0,
129
+ "num_inference_steps": 10,
130
+ "sigma_min": 0.0,
131
+ "num_integration_steps": 100,
132
+ "integration_method": "euler",
133
+ "timestep_sampling_strategy": "beta",
134
+ "timestep_sampling_s": 0.999,
135
+ "timestep_sampling_alpha": 1.5,
136
+ "timestep_sampling_beta": 1.0,
137
+ "hidden_dim": 512,
138
+ "num_layers": 6,
139
+ "num_heads": 8,
140
+ "dropout": 0.1,
141
+ "use_positional_encoding": false,
142
+ "timestep_embed_dim": 256,
143
+ "use_rope": true,
144
+ "rope_base": 10000.0,
145
+ "vision_encoder_name": "facebook/dinov3-vitb16-pretrain-lvd1689m",
146
+ "use_separate_rgb_encoder_per_camera": false,
147
+ "vision_encoder_lr_multiplier": 0.1,
148
+ "image_resize_shape": [
149
+ 224,
150
+ 304
151
+ ],
152
+ "image_crop_shape": null,
153
+ "image_crop_is_random": true,
154
+ "text_encoder_name": "openai/clip-vit-base-patch16",
155
+ "tokenizer_max_length": 77,
156
+ "tokenizer_padding": "max_length",
157
+ "tokenizer_padding_side": "right",
158
+ "tokenizer_truncation": true,
159
+ "normalization_mapping": {
160
+ "VISUAL": "MEAN_STD",
161
+ "STATE": "MIN_MAX",
162
+ "ACTION": "MIN_MAX"
163
+ },
164
+ "optimizer_lr": 2e-05,
165
+ "optimizer_betas": [
166
+ 0.95,
167
+ 0.999
168
+ ],
169
+ "optimizer_eps": 1e-08,
170
+ "optimizer_weight_decay": 0.0,
171
+ "scheduler_name": "cosine",
172
+ "scheduler_warmup_steps": 0,
173
+ "do_mask_loss_for_padding": false,
174
+ "drop_n_last_frames": 7
175
+ },
176
+ "reward_model": null,
177
+ "output_dir": "outputs/fold6_nocrop_rgray",
178
+ "job_name": "multi_task_dit",
179
+ "resume": false,
180
+ "seed": 1000,
181
+ "cudnn_deterministic": false,
182
+ "num_workers": 16,
183
+ "batch_size": 192,
184
+ "prefetch_factor": 4,
185
+ "persistent_workers": true,
186
+ "steps": 30000,
187
+ "eval_freq": 0,
188
+ "log_freq": 50,
189
+ "tolerance_s": 0.0001,
190
+ "save_checkpoint": true,
191
+ "save_freq": 2000,
192
+ "use_policy_training_preset": true,
193
+ "optimizer": {
194
+ "type": "adam",
195
+ "lr": 2e-05,
196
+ "weight_decay": 0.0,
197
+ "grad_clip_norm": 10.0,
198
+ "betas": [
199
+ 0.95,
200
+ 0.999
201
+ ],
202
+ "eps": 1e-08
203
+ },
204
+ "scheduler": {
205
+ "type": "diffuser",
206
+ "num_warmup_steps": 0,
207
+ "name": "cosine"
208
+ },
209
+ "eval": {
210
+ "n_episodes": 50,
211
+ "batch_size": 19,
212
+ "use_async_envs": true
213
+ },
214
+ "wandb": {
215
+ "enable": true,
216
+ "disable_artifact": false,
217
+ "project": "laundry_folding_robot",
218
+ "entity": null,
219
+ "notes": "run2_nocrop_rgray",
220
+ "run_id": "ir5slnwq",
221
+ "mode": null,
222
+ "add_tags": true
223
+ },
224
+ "peft": null,
225
+ "sample_weighting": null,
226
+ "rename_map": {},
227
+ "checkpoint_path": null
228
+ }