Hugging Face's logo

Codingpae
/
20261009_diff_100k

Safetensors
Model card Files
xet
 Community
Codingpae commited on
Commit
433e80c
·
verified ·
1 Parent(s): be4921d

Upload folder using huggingface_hub

Browse files
Files changed (7) hide show
  1. config.json +102 -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 +72 -0
  6. policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  7. train_config.json +255 -0
config.json ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "diffusion",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.images.headcam": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 480,
10
+ 640
11
+ ]
12
+ },
13
+ "observation.images.camera_01": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 480,
18
+ 640
19
+ ]
20
+ },
21
+ "observation.images.camera_02": {
22
+ "type": "VISUAL",
23
+ "shape": [
24
+ 3,
25
+ 480,
26
+ 640
27
+ ]
28
+ },
29
+ "observation.state": {
30
+ "type": "STATE",
31
+ "shape": [
32
+ 16
33
+ ]
34
+ }
35
+ },
36
+ "output_features": {
37
+ "action": {
38
+ "type": "ACTION",
39
+ "shape": [
40
+ 16
41
+ ]
42
+ }
43
+ },
44
+ "device": "cuda",
45
+ "use_amp": false,
46
+ "use_peft": false,
47
+ "push_to_hub": false,
48
+ "repo_id": null,
49
+ "private": null,
50
+ "tags": null,
51
+ "license": null,
52
+ "pretrained_path": null,
53
+ "pretrained_revision": null,
54
+ "horizon": 64,
55
+ "n_action_steps": 32,
56
+ "normalization_mapping": {
57
+ "VISUAL": "MEAN_STD",
58
+ "STATE": "MIN_MAX",
59
+ "ACTION": "MIN_MAX"
60
+ },
61
+ "drop_n_last_frames": 7,
62
+ "vision_backbone": "resnet18",
63
+ "resize_shape": null,
64
+ "crop_ratio": 1.0,
65
+ "crop_shape": null,
66
+ "crop_is_random": true,
67
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
68
+ "use_group_norm": false,
69
+ "spatial_softmax_num_keypoints": 32,
70
+ "use_separate_rgb_encoder_per_camera": true,
71
+ "down_dims": [
72
+ 512,
73
+ 1024,
74
+ 2048
75
+ ],
76
+ "kernel_size": 5,
77
+ "n_groups": 8,
78
+ "diffusion_step_embed_dim": 128,
79
+ "use_film_scale_modulation": true,
80
+ "gradient_checkpointing": false,
81
+ "noise_scheduler_type": "DDPM",
82
+ "num_train_timesteps": 100,
83
+ "beta_schedule": "squaredcos_cap_v2",
84
+ "beta_start": 0.0001,
85
+ "beta_end": 0.02,
86
+ "prediction_type": "epsilon",
87
+ "clip_sample": true,
88
+ "clip_sample_range": 1.0,
89
+ "num_inference_steps": null,
90
+ "compile_model": false,
91
+ "compile_mode": "reduce-overhead",
92
+ "do_mask_loss_for_padding": false,
93
+ "optimizer_lr": 0.0001,
94
+ "optimizer_betas": [
95
+ 0.95,
96
+ 0.999
97
+ ],
98
+ "optimizer_eps": 1e-08,
99
+ "optimizer_weight_decay": 1e-06,
100
+ "scheduler_name": "cosine",
101
+ "scheduler_warmup_steps": 500
102
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:548e85a9b765a911097e3ad1fdb588616eeec9a195b40be76fdb053aeacc5977
3
+ size 1173370056
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
+ 16
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:d89ca8f89b87524e79d1ae3bc10620993dc2be4714b70126b17fc95e29bc155b
3
+ size 9456
policy_preprocessor.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": "device_processor",
16
+ "config": {
17
+ "device": "cuda",
18
+ "float_dtype": null
19
+ }
20
+ },
21
+ {
22
+ "registry_name": "normalizer_processor",
23
+ "config": {
24
+ "eps": 1e-08,
25
+ "features": {
26
+ "observation.images.headcam": {
27
+ "type": "VISUAL",
28
+ "shape": [
29
+ 3,
30
+ 480,
31
+ 640
32
+ ]
33
+ },
34
+ "observation.images.camera_01": {
35
+ "type": "VISUAL",
36
+ "shape": [
37
+ 3,
38
+ 480,
39
+ 640
40
+ ]
41
+ },
42
+ "observation.images.camera_02": {
43
+ "type": "VISUAL",
44
+ "shape": [
45
+ 3,
46
+ 480,
47
+ 640
48
+ ]
49
+ },
50
+ "observation.state": {
51
+ "type": "STATE",
52
+ "shape": [
53
+ 16
54
+ ]
55
+ },
56
+ "action": {
57
+ "type": "ACTION",
58
+ "shape": [
59
+ 16
60
+ ]
61
+ }
62
+ },
63
+ "norm_map": {
64
+ "VISUAL": "MEAN_STD",
65
+ "STATE": "MIN_MAX",
66
+ "ACTION": "MIN_MAX"
67
+ }
68
+ },
69
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
70
+ }
71
+ ]
72
+ }
policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f57362ea36935bd9978a178a32ff83d00205edee94c6d1ff018ace75117290b
3
+ size 9464
train_config.json ADDED
@@ -0,0 +1,255 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "local/20261009",
4
+ "repo_type": "dataset",
5
+ "root": "/home/irteam/datasets/20261009",
6
+ "episodes": null,
7
+ "image_transforms": {
8
+ "enable": false,
9
+ "max_num_transforms": 3,
10
+ "random_order": false,
11
+ "tfs": {
12
+ "brightness": {
13
+ "weight": 1.0,
14
+ "type": "ColorJitter",
15
+ "kwargs": {
16
+ "brightness": [
17
+ 0.8,
18
+ 1.2
19
+ ]
20
+ }
21
+ },
22
+ "contrast": {
23
+ "weight": 1.0,
24
+ "type": "ColorJitter",
25
+ "kwargs": {
26
+ "contrast": [
27
+ 0.8,
28
+ 1.2
29
+ ]
30
+ }
31
+ },
32
+ "saturation": {
33
+ "weight": 1.0,
34
+ "type": "ColorJitter",
35
+ "kwargs": {
36
+ "saturation": [
37
+ 0.5,
38
+ 1.5
39
+ ]
40
+ }
41
+ },
42
+ "hue": {
43
+ "weight": 1.0,
44
+ "type": "ColorJitter",
45
+ "kwargs": {
46
+ "hue": [
47
+ -0.05,
48
+ 0.05
49
+ ]
50
+ }
51
+ },
52
+ "sharpness": {
53
+ "weight": 1.0,
54
+ "type": "SharpnessJitter",
55
+ "kwargs": {
56
+ "sharpness": [
57
+ 0.5,
58
+ 1.5
59
+ ]
60
+ }
61
+ },
62
+ "affine": {
63
+ "weight": 1.0,
64
+ "type": "RandomAffine",
65
+ "kwargs": {
66
+ "degrees": [
67
+ -5.0,
68
+ 5.0
69
+ ],
70
+ "translate": [
71
+ 0.05,
72
+ 0.05
73
+ ]
74
+ }
75
+ }
76
+ }
77
+ },
78
+ "revision": null,
79
+ "use_imagenet_stats": true,
80
+ "video_backend": "pyav",
81
+ "return_uint8": false,
82
+ "depth_output_unit": "mm",
83
+ "streaming": false,
84
+ "eval_split": 0.0
85
+ },
86
+ "env": null,
87
+ "policy": {
88
+ "type": "diffusion",
89
+ "n_obs_steps": 2,
90
+ "input_features": {
91
+ "observation.images.headcam": {
92
+ "type": "VISUAL",
93
+ "shape": [
94
+ 3,
95
+ 480,
96
+ 640
97
+ ]
98
+ },
99
+ "observation.images.camera_01": {
100
+ "type": "VISUAL",
101
+ "shape": [
102
+ 3,
103
+ 480,
104
+ 640
105
+ ]
106
+ },
107
+ "observation.images.camera_02": {
108
+ "type": "VISUAL",
109
+ "shape": [
110
+ 3,
111
+ 480,
112
+ 640
113
+ ]
114
+ },
115
+ "observation.state": {
116
+ "type": "STATE",
117
+ "shape": [
118
+ 16
119
+ ]
120
+ }
121
+ },
122
+ "output_features": {
123
+ "action": {
124
+ "type": "ACTION",
125
+ "shape": [
126
+ 16
127
+ ]
128
+ }
129
+ },
130
+ "device": "cuda",
131
+ "use_amp": false,
132
+ "use_peft": false,
133
+ "push_to_hub": false,
134
+ "repo_id": null,
135
+ "private": null,
136
+ "tags": null,
137
+ "license": null,
138
+ "pretrained_path": null,
139
+ "pretrained_revision": null,
140
+ "horizon": 64,
141
+ "n_action_steps": 32,
142
+ "normalization_mapping": {
143
+ "VISUAL": "MEAN_STD",
144
+ "STATE": "MIN_MAX",
145
+ "ACTION": "MIN_MAX"
146
+ },
147
+ "drop_n_last_frames": 7,
148
+ "vision_backbone": "resnet18",
149
+ "resize_shape": null,
150
+ "crop_ratio": 1.0,
151
+ "crop_shape": null,
152
+ "crop_is_random": true,
153
+ "pretrained_backbone_weights": "ResNet18_Weights.IMAGENET1K_V1",
154
+ "use_group_norm": false,
155
+ "spatial_softmax_num_keypoints": 32,
156
+ "use_separate_rgb_encoder_per_camera": true,
157
+ "down_dims": [
158
+ 512,
159
+ 1024,
160
+ 2048
161
+ ],
162
+ "kernel_size": 5,
163
+ "n_groups": 8,
164
+ "diffusion_step_embed_dim": 128,
165
+ "use_film_scale_modulation": true,
166
+ "gradient_checkpointing": false,
167
+ "noise_scheduler_type": "DDPM",
168
+ "num_train_timesteps": 100,
169
+ "beta_schedule": "squaredcos_cap_v2",
170
+ "beta_start": 0.0001,
171
+ "beta_end": 0.02,
172
+ "prediction_type": "epsilon",
173
+ "clip_sample": true,
174
+ "clip_sample_range": 1.0,
175
+ "num_inference_steps": null,
176
+ "compile_model": false,
177
+ "compile_mode": "reduce-overhead",
178
+ "do_mask_loss_for_padding": false,
179
+ "optimizer_lr": 0.0001,
180
+ "optimizer_betas": [
181
+ 0.95,
182
+ 0.999
183
+ ],
184
+ "optimizer_eps": 1e-08,
185
+ "optimizer_weight_decay": 1e-06,
186
+ "scheduler_name": "cosine",
187
+ "scheduler_warmup_steps": 500
188
+ },
189
+ "reward_model": null,
190
+ "output_dir": "/home/irteam/datasets/train/20261009_diff",
191
+ "job_name": "20261009_diff",
192
+ "resume": false,
193
+ "seed": 1000,
194
+ "cudnn_deterministic": false,
195
+ "num_workers": 8,
196
+ "batch_size": 64,
197
+ "prefetch_factor": 4,
198
+ "persistent_workers": true,
199
+ "dataloader_multiprocessing_context": "spawn",
200
+ "steps": 100000,
201
+ "env_eval_freq": 20000,
202
+ "log_freq": 200,
203
+ "eval_steps": 0,
204
+ "max_eval_samples": 0,
205
+ "tolerance_s": 0.0001,
206
+ "save_checkpoint": true,
207
+ "save_freq": 10000,
208
+ "use_policy_training_preset": true,
209
+ "optimizer": {
210
+ "type": "adam",
211
+ "lr": 0.0001,
212
+ "weight_decay": 1e-06,
213
+ "grad_clip_norm": 10.0,
214
+ "betas": [
215
+ 0.95,
216
+ 0.999
217
+ ],
218
+ "eps": 1e-08
219
+ },
220
+ "scheduler": {
221
+ "type": "diffuser",
222
+ "num_warmup_steps": 500,
223
+ "name": "cosine"
224
+ },
225
+ "eval": {
226
+ "n_episodes": 50,
227
+ "batch_size": 50,
228
+ "use_async_envs": true,
229
+ "recording": false,
230
+ "recording_repo_id": null,
231
+ "recording_private": false
232
+ },
233
+ "wandb": {
234
+ "enable": false,
235
+ "disable_artifact": false,
236
+ "project": "lerobot",
237
+ "entity": null,
238
+ "notes": null,
239
+ "run_id": null,
240
+ "mode": null,
241
+ "add_tags": true
242
+ },
243
+ "peft": null,
244
+ "job": {
245
+ "target": null,
246
+ "image": "huggingface/lerobot-gpu:latest",
247
+ "timeout": "2d",
248
+ "detach": false,
249
+ "tags": []
250
+ },
251
+ "save_checkpoint_to_hub": false,
252
+ "sample_weighting": null,
253
+ "rename_map": {},
254
+ "checkpoint_path": null
255
+ }