Hugging Face's logo

RzZ
/
cpolicy3-20k

Safetensors
Model card Files
xet
 Community
RzZ commited on
Commit
e4e0e80
·
verified ·
1 Parent(s): c8ce1e9

Upload folder using huggingface_hub

Browse files
Files changed (7) hide show
  1. config.json +128 -0
  2. model.safetensors +3 -0
  3. policy_postprocessor.json +38 -0
  4. policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  5. policy_preprocessor.json +103 -0
  6. policy_preprocessor_step_5_normalizer_processor.safetensors +3 -0
  7. train_config.json +264 -0
config.json ADDED
@@ -0,0 +1,128 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "multi_task_dit",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 26
9
+ ]
10
+ },
11
+ "observation.images.center_camera": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 256,
16
+ 256
17
+ ]
18
+ },
19
+ "observation.images.right_camera": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 256,
24
+ 256
25
+ ]
26
+ },
27
+ "observation.images.center_camera_center_crop": {
28
+ "type": "VISUAL",
29
+ "shape": [
30
+ 3,
31
+ 256,
32
+ 256
33
+ ]
34
+ }
35
+ },
36
+ "output_features": {
37
+ "action": {
38
+ "type": "ACTION",
39
+ "shape": [
40
+ 7
41
+ ]
42
+ }
43
+ },
44
+ "device": "cuda",
45
+ "use_amp": false,
46
+ "use_peft": false,
47
+ "push_to_hub": true,
48
+ "repo_id": "RzZ/cpolicy3",
49
+ "private": null,
50
+ "tags": null,
51
+ "license": null,
52
+ "pretrained_path": "RzZ/cpolicy-2-40k",
53
+ "horizon": 24,
54
+ "n_action_steps": 16,
55
+ "objective": "flow_matching",
56
+ "noise_scheduler_type": "DDPM",
57
+ "num_train_timesteps": 100,
58
+ "beta_schedule": "squaredcos_cap_v2",
59
+ "beta_start": 0.0001,
60
+ "beta_end": 0.02,
61
+ "prediction_type": "epsilon",
62
+ "clip_sample": true,
63
+ "clip_sample_range": 1.0,
64
+ "num_inference_steps": null,
65
+ "sigma_min": 0.0,
66
+ "num_integration_steps": 100,
67
+ "integration_method": "euler",
68
+ "timestep_sampling_strategy": "beta",
69
+ "timestep_sampling_s": 0.999,
70
+ "timestep_sampling_alpha": 1.5,
71
+ "timestep_sampling_beta": 1.0,
72
+ "use_relative_actions": true,
73
+ "relative_exclude_joints": [
74
+ "gripper"
75
+ ],
76
+ "action_feature_names": [
77
+ "position.x",
78
+ "position.y",
79
+ "position.z",
80
+ "orientation.x",
81
+ "orientation.y",
82
+ "orientation.z",
83
+ "orientation.w"
84
+ ],
85
+ "hidden_dim": 512,
86
+ "num_layers": 6,
87
+ "num_heads": 8,
88
+ "dropout": 0.1,
89
+ "use_positional_encoding": false,
90
+ "timestep_embed_dim": 256,
91
+ "use_rope": true,
92
+ "rope_base": 10000.0,
93
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
94
+ "use_separate_rgb_encoder_per_camera": false,
95
+ "vision_encoder_lr_multiplier": 0.1,
96
+ "image_resize_shape": [
97
+ 224,
98
+ 224
99
+ ],
100
+ "image_crop_shape": [
101
+ 224,
102
+ 224
103
+ ],
104
+ "image_crop_is_random": true,
105
+ "freeze_vision_encoder": false,
106
+ "text_encoder_name": "openai/clip-vit-base-patch16",
107
+ "tokenizer_max_length": 77,
108
+ "tokenizer_padding": "max_length",
109
+ "tokenizer_padding_side": "right",
110
+ "tokenizer_truncation": true,
111
+ "text_embed_table": false,
112
+ "normalization_mapping": {
113
+ "VISUAL": "MEAN_STD",
114
+ "STATE": "MIN_MAX",
115
+ "ACTION": "MIN_MAX"
116
+ },
117
+ "optimizer_lr": 2e-05,
118
+ "optimizer_betas": [
119
+ 0.95,
120
+ 0.999
121
+ ],
122
+ "optimizer_eps": 1e-08,
123
+ "optimizer_weight_decay": 0.0,
124
+ "scheduler_name": "cosine",
125
+ "scheduler_warmup_steps": 0,
126
+ "do_mask_loss_for_padding": false,
127
+ "drop_n_last_frames": 8
128
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:68a11e68428ffbad493364d8752c9957241a3f21b3d96e2e76cc5d12e7278fcd
3
+ size 902158548
policy_postprocessor.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "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": "absolute_actions_processor",
26
+ "config": {
27
+ "enabled": true
28
+ }
29
+ },
30
+ {
31
+ "registry_name": "device_processor",
32
+ "config": {
33
+ "device": "cpu",
34
+ "float_dtype": null
35
+ }
36
+ }
37
+ ]
38
+ }
policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b34f6a69853e31a9f3c0de7625a69d72bc41d9052b1324613785b8ce86d845d9
3
+ size 9716
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": "delta_actions_processor",
16
+ "config": {
17
+ "enabled": true,
18
+ "exclude_joints": [
19
+ "gripper"
20
+ ],
21
+ "action_names": [
22
+ "position.x",
23
+ "position.y",
24
+ "position.z",
25
+ "orientation.x",
26
+ "orientation.y",
27
+ "orientation.z",
28
+ "orientation.w"
29
+ ],
30
+ "convert_relative_quat_to_rotvec": true,
31
+ "use_ee_frame_pose": true
32
+ }
33
+ },
34
+ {
35
+ "registry_name": "tokenizer_processor",
36
+ "config": {
37
+ "max_length": 77,
38
+ "task_key": "task",
39
+ "padding_side": "right",
40
+ "padding": "max_length",
41
+ "truncation": true,
42
+ "tokenizer_name": "openai/clip-vit-base-patch16"
43
+ }
44
+ },
45
+ {
46
+ "registry_name": "device_processor",
47
+ "config": {
48
+ "device": "cuda",
49
+ "float_dtype": null
50
+ }
51
+ },
52
+ {
53
+ "registry_name": "normalizer_processor",
54
+ "config": {
55
+ "eps": 1e-08,
56
+ "features": {
57
+ "observation.state": {
58
+ "type": "STATE",
59
+ "shape": [
60
+ 26
61
+ ]
62
+ },
63
+ "observation.images.center_camera": {
64
+ "type": "VISUAL",
65
+ "shape": [
66
+ 3,
67
+ 256,
68
+ 256
69
+ ]
70
+ },
71
+ "observation.images.right_camera": {
72
+ "type": "VISUAL",
73
+ "shape": [
74
+ 3,
75
+ 256,
76
+ 256
77
+ ]
78
+ },
79
+ "observation.images.center_camera_center_crop": {
80
+ "type": "VISUAL",
81
+ "shape": [
82
+ 3,
83
+ 256,
84
+ 256
85
+ ]
86
+ },
87
+ "action": {
88
+ "type": "ACTION",
89
+ "shape": [
90
+ 7
91
+ ]
92
+ }
93
+ },
94
+ "norm_map": {
95
+ "VISUAL": "MEAN_STD",
96
+ "STATE": "MIN_MAX",
97
+ "ACTION": "MIN_MAX"
98
+ }
99
+ },
100
+ "state_file": "policy_preprocessor_step_5_normalizer_processor.safetensors"
101
+ }
102
+ ]
103
+ }
policy_preprocessor_step_5_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b34f6a69853e31a9f3c0de7625a69d72bc41d9052b1324613785b8ce86d845d9
3
+ size 9716
train_config.json ADDED
@@ -0,0 +1,264 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "RzZ/t1-t2-isaac-v7-edit",
4
+ "root": "",
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": null,
78
+ "use_imagenet_stats": true,
79
+ "video_backend": "torchcodec",
80
+ "streaming": false
81
+ },
82
+ "env": null,
83
+ "policy": {
84
+ "type": "multi_task_dit",
85
+ "n_obs_steps": 1,
86
+ "input_features": {
87
+ "observation.state": {
88
+ "type": "STATE",
89
+ "shape": [
90
+ 26
91
+ ]
92
+ },
93
+ "observation.images.center_camera": {
94
+ "type": "VISUAL",
95
+ "shape": [
96
+ 3,
97
+ 256,
98
+ 256
99
+ ]
100
+ },
101
+ "observation.images.right_camera": {
102
+ "type": "VISUAL",
103
+ "shape": [
104
+ 3,
105
+ 256,
106
+ 256
107
+ ]
108
+ },
109
+ "observation.images.center_camera_center_crop": {
110
+ "type": "VISUAL",
111
+ "shape": [
112
+ 3,
113
+ 256,
114
+ 256
115
+ ]
116
+ }
117
+ },
118
+ "output_features": {
119
+ "action": {
120
+ "type": "ACTION",
121
+ "shape": [
122
+ 7
123
+ ]
124
+ }
125
+ },
126
+ "device": "cuda",
127
+ "use_amp": false,
128
+ "use_peft": false,
129
+ "push_to_hub": true,
130
+ "repo_id": "RzZ/cpolicy3",
131
+ "private": null,
132
+ "tags": null,
133
+ "license": null,
134
+ "pretrained_path": "RzZ/cpolicy-2-40k",
135
+ "horizon": 24,
136
+ "n_action_steps": 16,
137
+ "objective": "flow_matching",
138
+ "noise_scheduler_type": "DDPM",
139
+ "num_train_timesteps": 100,
140
+ "beta_schedule": "squaredcos_cap_v2",
141
+ "beta_start": 0.0001,
142
+ "beta_end": 0.02,
143
+ "prediction_type": "epsilon",
144
+ "clip_sample": true,
145
+ "clip_sample_range": 1.0,
146
+ "num_inference_steps": null,
147
+ "sigma_min": 0.0,
148
+ "num_integration_steps": 100,
149
+ "integration_method": "euler",
150
+ "timestep_sampling_strategy": "beta",
151
+ "timestep_sampling_s": 0.999,
152
+ "timestep_sampling_alpha": 1.5,
153
+ "timestep_sampling_beta": 1.0,
154
+ "use_relative_actions": true,
155
+ "relative_exclude_joints": [
156
+ "gripper"
157
+ ],
158
+ "action_feature_names": [
159
+ "position.x",
160
+ "position.y",
161
+ "position.z",
162
+ "orientation.x",
163
+ "orientation.y",
164
+ "orientation.z",
165
+ "orientation.w"
166
+ ],
167
+ "hidden_dim": 512,
168
+ "num_layers": 6,
169
+ "num_heads": 8,
170
+ "dropout": 0.1,
171
+ "use_positional_encoding": false,
172
+ "timestep_embed_dim": 256,
173
+ "use_rope": true,
174
+ "rope_base": 10000.0,
175
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
176
+ "use_separate_rgb_encoder_per_camera": false,
177
+ "vision_encoder_lr_multiplier": 0.1,
178
+ "image_resize_shape": [
179
+ 224,
180
+ 224
181
+ ],
182
+ "image_crop_shape": [
183
+ 224,
184
+ 224
185
+ ],
186
+ "image_crop_is_random": true,
187
+ "freeze_vision_encoder": false,
188
+ "text_encoder_name": "openai/clip-vit-base-patch16",
189
+ "tokenizer_max_length": 77,
190
+ "tokenizer_padding": "max_length",
191
+ "tokenizer_padding_side": "right",
192
+ "tokenizer_truncation": true,
193
+ "text_embed_table": false,
194
+ "normalization_mapping": {
195
+ "VISUAL": "MEAN_STD",
196
+ "STATE": "MIN_MAX",
197
+ "ACTION": "MIN_MAX"
198
+ },
199
+ "optimizer_lr": 2e-05,
200
+ "optimizer_betas": [
201
+ 0.95,
202
+ 0.999
203
+ ],
204
+ "optimizer_eps": 1e-08,
205
+ "optimizer_weight_decay": 0.0,
206
+ "scheduler_name": "cosine",
207
+ "scheduler_warmup_steps": 0,
208
+ "do_mask_loss_for_padding": false,
209
+ "drop_n_last_frames": 8
210
+ },
211
+ "output_dir": "/workspace/mtd-checkpoint-3",
212
+ "job_name": "multi_task_dit",
213
+ "resume": false,
214
+ "seed": 1000,
215
+ "cudnn_deterministic": false,
216
+ "num_workers": 4,
217
+ "batch_size": 64,
218
+ "steps": 50000,
219
+ "eval_freq": 20000,
220
+ "log_freq": 50,
221
+ "tolerance_s": 0.0001,
222
+ "save_checkpoint": true,
223
+ "save_freq": 10000,
224
+ "use_policy_training_preset": true,
225
+ "optimizer": {
226
+ "type": "adam",
227
+ "lr": 2e-05,
228
+ "weight_decay": 0.0,
229
+ "grad_clip_norm": 10.0,
230
+ "betas": [
231
+ 0.95,
232
+ 0.999
233
+ ],
234
+ "eps": 1e-08
235
+ },
236
+ "scheduler": {
237
+ "type": "diffuser",
238
+ "num_warmup_steps": 0,
239
+ "name": "cosine"
240
+ },
241
+ "eval": {
242
+ "n_episodes": 50,
243
+ "batch_size": 50,
244
+ "use_async_envs": false
245
+ },
246
+ "wandb": {
247
+ "enable": true,
248
+ "disable_artifact": true,
249
+ "project": "lerobot",
250
+ "entity": null,
251
+ "notes": null,
252
+ "run_id": "jbq6ukzu",
253
+ "mode": null,
254
+ "add_tags": true
255
+ },
256
+ "peft": null,
257
+ "use_rabc": false,
258
+ "rabc_progress_path": null,
259
+ "rabc_kappa": 0.01,
260
+ "rabc_epsilon": 1e-06,
261
+ "rabc_head_mode": "sparse",
262
+ "rename_map": {},
263
+ "checkpoint_path": null
264
+ }