bcsp commited on
Commit
8a79fa7
·
verified ·
1 Parent(s): 0d8e6f3

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. checkpoints/010000/pretrained_model/config.json +121 -0
  2. checkpoints/010000/pretrained_model/model.safetensors +3 -0
  3. checkpoints/010000/pretrained_model/policy_postprocessor.json +32 -0
  4. checkpoints/010000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  5. checkpoints/010000/pretrained_model/policy_preprocessor.json +78 -0
  6. checkpoints/010000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  7. checkpoints/010000/pretrained_model/train_config.json +608 -0
  8. checkpoints/010000/training_state/optimizer_param_groups.json +308 -0
  9. checkpoints/010000/training_state/optimizer_state.safetensors +3 -0
  10. checkpoints/010000/training_state/rng_state.safetensors +3 -0
  11. checkpoints/010000/training_state/scheduler_state.json +17 -0
  12. checkpoints/010000/training_state/training_step.json +3 -0
  13. checkpoints/020000/pretrained_model/config.json +121 -0
  14. checkpoints/020000/pretrained_model/model.safetensors +3 -0
  15. checkpoints/020000/pretrained_model/policy_postprocessor.json +32 -0
  16. checkpoints/020000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  17. checkpoints/020000/pretrained_model/policy_preprocessor.json +78 -0
  18. checkpoints/020000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  19. checkpoints/020000/pretrained_model/train_config.json +608 -0
  20. checkpoints/020000/training_state/optimizer_param_groups.json +308 -0
  21. checkpoints/020000/training_state/optimizer_state.safetensors +3 -0
  22. checkpoints/020000/training_state/rng_state.safetensors +3 -0
  23. checkpoints/020000/training_state/scheduler_state.json +17 -0
  24. checkpoints/020000/training_state/training_step.json +3 -0
  25. checkpoints/030000/pretrained_model/config.json +121 -0
  26. checkpoints/030000/pretrained_model/model.safetensors +3 -0
  27. checkpoints/030000/pretrained_model/policy_postprocessor.json +32 -0
  28. checkpoints/030000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  29. checkpoints/030000/pretrained_model/policy_preprocessor.json +78 -0
  30. checkpoints/030000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  31. checkpoints/030000/pretrained_model/train_config.json +608 -0
  32. checkpoints/030000/training_state/optimizer_param_groups.json +308 -0
  33. checkpoints/030000/training_state/optimizer_state.safetensors +3 -0
  34. checkpoints/030000/training_state/rng_state.safetensors +3 -0
  35. checkpoints/030000/training_state/scheduler_state.json +17 -0
  36. checkpoints/030000/training_state/training_step.json +3 -0
  37. checkpoints/040000/pretrained_model/config.json +121 -0
  38. checkpoints/040000/pretrained_model/model.safetensors +3 -0
  39. checkpoints/040000/pretrained_model/policy_postprocessor.json +32 -0
  40. checkpoints/040000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  41. checkpoints/040000/pretrained_model/policy_preprocessor.json +78 -0
  42. checkpoints/040000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  43. checkpoints/040000/pretrained_model/train_config.json +608 -0
  44. checkpoints/040000/training_state/optimizer_param_groups.json +308 -0
  45. checkpoints/040000/training_state/optimizer_state.safetensors +3 -0
  46. checkpoints/040000/training_state/rng_state.safetensors +3 -0
  47. checkpoints/040000/training_state/scheduler_state.json +17 -0
  48. checkpoints/040000/training_state/training_step.json +3 -0
  49. checkpoints/050000/pretrained_model/config.json +121 -0
  50. checkpoints/050000/pretrained_model/model.safetensors +3 -0
checkpoints/010000/pretrained_model/config.json ADDED
@@ -0,0 +1,121 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "aic_multi_task_dit",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.images.left_camera": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 256,
10
+ 288
11
+ ]
12
+ },
13
+ "observation.images.center_camera": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 256,
18
+ 288
19
+ ]
20
+ },
21
+ "observation.images.right_camera": {
22
+ "type": "VISUAL",
23
+ "shape": [
24
+ 3,
25
+ 256,
26
+ 288
27
+ ]
28
+ },
29
+ "observation.state": {
30
+ "type": "STATE",
31
+ "shape": [
32
+ 26
33
+ ]
34
+ },
35
+ "observation.wrench": {
36
+ "type": "STATE",
37
+ "shape": [
38
+ 6
39
+ ]
40
+ }
41
+ },
42
+ "output_features": {
43
+ "action": {
44
+ "type": "ACTION",
45
+ "shape": [
46
+ 7
47
+ ]
48
+ }
49
+ },
50
+ "device": "cuda",
51
+ "use_amp": false,
52
+ "use_peft": false,
53
+ "push_to_hub": false,
54
+ "repo_id": null,
55
+ "private": null,
56
+ "tags": null,
57
+ "license": null,
58
+ "pretrained_path": null,
59
+ "horizon": 32,
60
+ "n_action_steps": 24,
61
+ "task_conditioning": "embedding",
62
+ "num_tasks": 10,
63
+ "objective": "diffusion",
64
+ "noise_scheduler_type": "DDPM",
65
+ "num_train_timesteps": 100,
66
+ "beta_schedule": "squaredcos_cap_v2",
67
+ "beta_start": 0.0001,
68
+ "beta_end": 0.02,
69
+ "prediction_type": "epsilon",
70
+ "clip_sample": true,
71
+ "clip_sample_range": 1.0,
72
+ "num_inference_steps": null,
73
+ "sigma_min": 0.0,
74
+ "num_integration_steps": 100,
75
+ "integration_method": "euler",
76
+ "timestep_sampling_strategy": "beta",
77
+ "timestep_sampling_s": 0.999,
78
+ "timestep_sampling_alpha": 1.5,
79
+ "timestep_sampling_beta": 1.0,
80
+ "hidden_dim": 512,
81
+ "num_layers": 6,
82
+ "num_heads": 8,
83
+ "dropout": 0.1,
84
+ "use_positional_encoding": false,
85
+ "timestep_embed_dim": 256,
86
+ "use_rope": true,
87
+ "rope_base": 10000.0,
88
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
89
+ "use_separate_rgb_encoder_per_camera": false,
90
+ "vision_encoder_lr_multiplier": 0.1,
91
+ "image_resize_shape": [
92
+ 232,
93
+ 232
94
+ ],
95
+ "image_crop_shape": [
96
+ 224,
97
+ 224
98
+ ],
99
+ "image_crop_is_random": true,
100
+ "text_encoder_name": "openai/clip-vit-base-patch16",
101
+ "tokenizer_max_length": 77,
102
+ "tokenizer_padding": "max_length",
103
+ "tokenizer_padding_side": "right",
104
+ "tokenizer_truncation": true,
105
+ "normalization_mapping": {
106
+ "VISUAL": "MEAN_STD",
107
+ "STATE": "MIN_MAX",
108
+ "ACTION": "MIN_MAX"
109
+ },
110
+ "optimizer_lr": 5e-05,
111
+ "optimizer_betas": [
112
+ 0.95,
113
+ 0.999
114
+ ],
115
+ "optimizer_eps": 1e-08,
116
+ "optimizer_weight_decay": 0.0001,
117
+ "scheduler_name": "cosine",
118
+ "scheduler_warmup_steps": 0,
119
+ "do_mask_loss_for_padding": false,
120
+ "drop_n_last_frames": 7
121
+ }
checkpoints/010000/pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1177383edcec91362235ba5b2bb991922375c07e459c554e9956766a52df11a2
3
+ size 857967220
checkpoints/010000/pretrained_model/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": "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
+ }
checkpoints/010000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d58fa83446e4b0bdfc09a02565d49fca3dd894c73486df2d1ce3a22e752cd6
3
+ size 11368
checkpoints/010000/pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.left_camera": {
27
+ "type": "VISUAL",
28
+ "shape": [
29
+ 3,
30
+ 256,
31
+ 288
32
+ ]
33
+ },
34
+ "observation.images.center_camera": {
35
+ "type": "VISUAL",
36
+ "shape": [
37
+ 3,
38
+ 256,
39
+ 288
40
+ ]
41
+ },
42
+ "observation.images.right_camera": {
43
+ "type": "VISUAL",
44
+ "shape": [
45
+ 3,
46
+ 256,
47
+ 288
48
+ ]
49
+ },
50
+ "observation.state": {
51
+ "type": "STATE",
52
+ "shape": [
53
+ 26
54
+ ]
55
+ },
56
+ "observation.wrench": {
57
+ "type": "STATE",
58
+ "shape": [
59
+ 6
60
+ ]
61
+ },
62
+ "action": {
63
+ "type": "ACTION",
64
+ "shape": [
65
+ 7
66
+ ]
67
+ }
68
+ },
69
+ "norm_map": {
70
+ "VISUAL": "MEAN_STD",
71
+ "STATE": "MIN_MAX",
72
+ "ACTION": "MIN_MAX"
73
+ }
74
+ },
75
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
76
+ }
77
+ ]
78
+ }
checkpoints/010000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d58fa83446e4b0bdfc09a02565d49fca3dd894c73486df2d1ce3a22e752cd6
3
+ size 11368
checkpoints/010000/pretrained_model/train_config.json ADDED
@@ -0,0 +1,608 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "local/local_aic_cheatcode_demos_dit_snap",
4
+ "root": "/opt/dlami/nvme/lerobot/local_aic_cheatcode_demos_dit_snap",
5
+ "episodes": [
6
+ 3,
7
+ 5,
8
+ 8,
9
+ 157,
10
+ 183,
11
+ 184,
12
+ 185,
13
+ 186,
14
+ 187,
15
+ 188,
16
+ 189,
17
+ 190,
18
+ 192,
19
+ 193,
20
+ 194,
21
+ 195,
22
+ 196,
23
+ 198,
24
+ 199,
25
+ 200,
26
+ 201,
27
+ 203,
28
+ 204,
29
+ 205,
30
+ 206,
31
+ 207,
32
+ 208,
33
+ 209,
34
+ 210,
35
+ 211,
36
+ 213,
37
+ 214,
38
+ 215,
39
+ 216,
40
+ 217,
41
+ 218,
42
+ 219,
43
+ 220,
44
+ 221,
45
+ 222,
46
+ 223,
47
+ 224,
48
+ 225,
49
+ 227,
50
+ 228,
51
+ 229,
52
+ 230,
53
+ 232,
54
+ 233,
55
+ 234,
56
+ 236,
57
+ 238,
58
+ 239,
59
+ 241,
60
+ 242,
61
+ 243,
62
+ 244,
63
+ 245,
64
+ 246,
65
+ 247,
66
+ 248,
67
+ 249,
68
+ 250,
69
+ 251,
70
+ 252,
71
+ 253,
72
+ 255,
73
+ 256,
74
+ 257,
75
+ 258,
76
+ 259,
77
+ 260,
78
+ 262,
79
+ 263,
80
+ 264,
81
+ 265,
82
+ 266,
83
+ 267,
84
+ 268,
85
+ 269,
86
+ 270,
87
+ 271,
88
+ 272,
89
+ 273,
90
+ 274,
91
+ 276,
92
+ 277,
93
+ 278,
94
+ 279,
95
+ 280,
96
+ 281,
97
+ 283,
98
+ 284,
99
+ 285,
100
+ 286,
101
+ 287,
102
+ 288,
103
+ 289,
104
+ 290,
105
+ 291,
106
+ 292,
107
+ 293,
108
+ 294,
109
+ 295,
110
+ 298,
111
+ 299,
112
+ 300,
113
+ 301,
114
+ 302,
115
+ 303,
116
+ 304,
117
+ 305,
118
+ 306,
119
+ 307,
120
+ 308,
121
+ 309,
122
+ 310,
123
+ 311,
124
+ 312,
125
+ 313,
126
+ 314,
127
+ 315,
128
+ 316,
129
+ 317,
130
+ 318,
131
+ 319,
132
+ 321,
133
+ 322,
134
+ 323,
135
+ 325,
136
+ 327,
137
+ 328,
138
+ 329,
139
+ 330,
140
+ 332,
141
+ 333,
142
+ 334,
143
+ 335,
144
+ 336,
145
+ 337,
146
+ 338,
147
+ 340,
148
+ 341,
149
+ 342,
150
+ 343,
151
+ 344,
152
+ 345,
153
+ 346,
154
+ 347,
155
+ 348,
156
+ 350,
157
+ 351,
158
+ 352,
159
+ 354,
160
+ 355,
161
+ 356,
162
+ 358,
163
+ 359,
164
+ 360,
165
+ 361,
166
+ 362,
167
+ 363,
168
+ 364,
169
+ 366,
170
+ 367,
171
+ 368,
172
+ 369,
173
+ 370,
174
+ 371,
175
+ 372,
176
+ 373,
177
+ 374,
178
+ 375,
179
+ 377,
180
+ 378,
181
+ 379,
182
+ 380,
183
+ 381,
184
+ 384,
185
+ 385,
186
+ 386,
187
+ 387,
188
+ 388,
189
+ 389,
190
+ 390,
191
+ 391,
192
+ 393,
193
+ 395,
194
+ 397,
195
+ 399,
196
+ 400,
197
+ 401,
198
+ 402,
199
+ 404,
200
+ 405,
201
+ 407,
202
+ 408,
203
+ 409,
204
+ 410,
205
+ 411,
206
+ 413,
207
+ 414,
208
+ 415,
209
+ 416,
210
+ 417,
211
+ 419,
212
+ 420,
213
+ 421,
214
+ 422,
215
+ 423,
216
+ 425,
217
+ 430,
218
+ 431,
219
+ 433,
220
+ 434,
221
+ 436,
222
+ 437,
223
+ 439,
224
+ 440,
225
+ 442,
226
+ 444,
227
+ 447,
228
+ 449,
229
+ 450,
230
+ 452,
231
+ 453,
232
+ 454,
233
+ 455,
234
+ 457,
235
+ 458,
236
+ 460,
237
+ 462,
238
+ 463,
239
+ 464,
240
+ 465,
241
+ 466,
242
+ 467,
243
+ 470,
244
+ 471,
245
+ 473,
246
+ 474,
247
+ 475,
248
+ 476,
249
+ 478,
250
+ 479,
251
+ 480,
252
+ 482,
253
+ 485,
254
+ 490,
255
+ 491,
256
+ 492,
257
+ 497,
258
+ 498,
259
+ 500,
260
+ 501,
261
+ 502,
262
+ 504,
263
+ 507,
264
+ 508,
265
+ 509,
266
+ 512,
267
+ 514,
268
+ 515,
269
+ 516,
270
+ 517,
271
+ 518,
272
+ 520,
273
+ 522,
274
+ 524,
275
+ 527,
276
+ 529,
277
+ 530,
278
+ 532,
279
+ 533,
280
+ 534,
281
+ 535,
282
+ 537,
283
+ 538,
284
+ 539,
285
+ 540,
286
+ 541,
287
+ 542,
288
+ 543,
289
+ 544,
290
+ 546,
291
+ 548,
292
+ 549,
293
+ 550,
294
+ 551,
295
+ 553,
296
+ 555,
297
+ 556,
298
+ 557,
299
+ 558,
300
+ 559,
301
+ 560,
302
+ 562,
303
+ 563,
304
+ 565,
305
+ 566,
306
+ 569,
307
+ 570,
308
+ 571,
309
+ 572,
310
+ 575,
311
+ 576,
312
+ 577,
313
+ 578,
314
+ 580,
315
+ 581,
316
+ 582,
317
+ 584,
318
+ 585,
319
+ 586,
320
+ 587,
321
+ 588,
322
+ 589,
323
+ 590,
324
+ 591,
325
+ 593,
326
+ 594,
327
+ 595,
328
+ 596,
329
+ 598,
330
+ 600,
331
+ 601,
332
+ 602,
333
+ 603,
334
+ 606,
335
+ 607,
336
+ 608,
337
+ 609,
338
+ 612,
339
+ 613,
340
+ 615,
341
+ 616,
342
+ 617,
343
+ 619,
344
+ 620,
345
+ 621,
346
+ 623,
347
+ 624,
348
+ 627,
349
+ 628,
350
+ 631,
351
+ 632,
352
+ 633,
353
+ 634,
354
+ 635,
355
+ 636
356
+ ],
357
+ "image_transforms": {
358
+ "enable": true,
359
+ "max_num_transforms": 3,
360
+ "random_order": true,
361
+ "tfs": {
362
+ "brightness": {
363
+ "weight": 1.0,
364
+ "type": "ColorJitter",
365
+ "kwargs": {
366
+ "brightness": [
367
+ 0.9,
368
+ 1.1
369
+ ]
370
+ }
371
+ },
372
+ "contrast": {
373
+ "weight": 1.0,
374
+ "type": "ColorJitter",
375
+ "kwargs": {
376
+ "contrast": [
377
+ 0.9,
378
+ 1.1
379
+ ]
380
+ }
381
+ },
382
+ "saturation": {
383
+ "weight": 1.0,
384
+ "type": "ColorJitter",
385
+ "kwargs": {
386
+ "saturation": [
387
+ 0.95,
388
+ 1.05
389
+ ]
390
+ }
391
+ },
392
+ "hue": {
393
+ "weight": 1.0,
394
+ "type": "ColorJitter",
395
+ "kwargs": {
396
+ "hue": [
397
+ -0.02,
398
+ 0.02
399
+ ]
400
+ }
401
+ },
402
+ "sharpness": {
403
+ "weight": 1.0,
404
+ "type": "SharpnessJitter",
405
+ "kwargs": {
406
+ "sharpness": [
407
+ 0.9,
408
+ 1.1
409
+ ]
410
+ }
411
+ },
412
+ "affine": {
413
+ "weight": 0.0,
414
+ "type": "RandomAffine",
415
+ "kwargs": {
416
+ "degrees": [
417
+ -5.0,
418
+ 5.0
419
+ ],
420
+ "translate": [
421
+ 0.05,
422
+ 0.05
423
+ ]
424
+ }
425
+ }
426
+ }
427
+ },
428
+ "revision": null,
429
+ "use_imagenet_stats": true,
430
+ "video_backend": "pyav",
431
+ "streaming": false
432
+ },
433
+ "env": null,
434
+ "policy": {
435
+ "type": "aic_multi_task_dit",
436
+ "n_obs_steps": 2,
437
+ "input_features": {
438
+ "observation.images.left_camera": {
439
+ "type": "VISUAL",
440
+ "shape": [
441
+ 3,
442
+ 256,
443
+ 288
444
+ ]
445
+ },
446
+ "observation.images.center_camera": {
447
+ "type": "VISUAL",
448
+ "shape": [
449
+ 3,
450
+ 256,
451
+ 288
452
+ ]
453
+ },
454
+ "observation.images.right_camera": {
455
+ "type": "VISUAL",
456
+ "shape": [
457
+ 3,
458
+ 256,
459
+ 288
460
+ ]
461
+ },
462
+ "observation.state": {
463
+ "type": "STATE",
464
+ "shape": [
465
+ 26
466
+ ]
467
+ },
468
+ "observation.wrench": {
469
+ "type": "STATE",
470
+ "shape": [
471
+ 6
472
+ ]
473
+ }
474
+ },
475
+ "output_features": {
476
+ "action": {
477
+ "type": "ACTION",
478
+ "shape": [
479
+ 7
480
+ ]
481
+ }
482
+ },
483
+ "device": "cuda",
484
+ "use_amp": false,
485
+ "use_peft": false,
486
+ "push_to_hub": false,
487
+ "repo_id": null,
488
+ "private": null,
489
+ "tags": null,
490
+ "license": null,
491
+ "pretrained_path": null,
492
+ "horizon": 32,
493
+ "n_action_steps": 24,
494
+ "task_conditioning": "embedding",
495
+ "num_tasks": 10,
496
+ "objective": "diffusion",
497
+ "noise_scheduler_type": "DDPM",
498
+ "num_train_timesteps": 100,
499
+ "beta_schedule": "squaredcos_cap_v2",
500
+ "beta_start": 0.0001,
501
+ "beta_end": 0.02,
502
+ "prediction_type": "epsilon",
503
+ "clip_sample": true,
504
+ "clip_sample_range": 1.0,
505
+ "num_inference_steps": null,
506
+ "sigma_min": 0.0,
507
+ "num_integration_steps": 100,
508
+ "integration_method": "euler",
509
+ "timestep_sampling_strategy": "beta",
510
+ "timestep_sampling_s": 0.999,
511
+ "timestep_sampling_alpha": 1.5,
512
+ "timestep_sampling_beta": 1.0,
513
+ "hidden_dim": 512,
514
+ "num_layers": 6,
515
+ "num_heads": 8,
516
+ "dropout": 0.1,
517
+ "use_positional_encoding": false,
518
+ "timestep_embed_dim": 256,
519
+ "use_rope": true,
520
+ "rope_base": 10000.0,
521
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
522
+ "use_separate_rgb_encoder_per_camera": false,
523
+ "vision_encoder_lr_multiplier": 0.1,
524
+ "image_resize_shape": [
525
+ 232,
526
+ 232
527
+ ],
528
+ "image_crop_shape": [
529
+ 224,
530
+ 224
531
+ ],
532
+ "image_crop_is_random": true,
533
+ "text_encoder_name": "openai/clip-vit-base-patch16",
534
+ "tokenizer_max_length": 77,
535
+ "tokenizer_padding": "max_length",
536
+ "tokenizer_padding_side": "right",
537
+ "tokenizer_truncation": true,
538
+ "normalization_mapping": {
539
+ "VISUAL": "MEAN_STD",
540
+ "STATE": "MIN_MAX",
541
+ "ACTION": "MIN_MAX"
542
+ },
543
+ "optimizer_lr": 5e-05,
544
+ "optimizer_betas": [
545
+ 0.95,
546
+ 0.999
547
+ ],
548
+ "optimizer_eps": 1e-08,
549
+ "optimizer_weight_decay": 0.0001,
550
+ "scheduler_name": "cosine",
551
+ "scheduler_warmup_steps": 0,
552
+ "do_mask_loss_for_padding": false,
553
+ "drop_n_last_frames": 7
554
+ },
555
+ "output_dir": "/opt/dlami/nvme/outputs/dit_diffusion_20260511_105440",
556
+ "job_name": "aic_multi_task_dit",
557
+ "resume": false,
558
+ "seed": 42,
559
+ "cudnn_deterministic": false,
560
+ "num_workers": 4,
561
+ "batch_size": 12,
562
+ "steps": 100000,
563
+ "eval_freq": 0,
564
+ "log_freq": 100,
565
+ "tolerance_s": 0.1,
566
+ "save_checkpoint": true,
567
+ "save_freq": 10000,
568
+ "use_policy_training_preset": true,
569
+ "optimizer": {
570
+ "type": "adam",
571
+ "lr": 5e-05,
572
+ "weight_decay": 0.0001,
573
+ "grad_clip_norm": 10.0,
574
+ "betas": [
575
+ 0.95,
576
+ 0.999
577
+ ],
578
+ "eps": 1e-08
579
+ },
580
+ "scheduler": {
581
+ "type": "diffuser",
582
+ "num_warmup_steps": 0,
583
+ "name": "cosine"
584
+ },
585
+ "eval": {
586
+ "n_episodes": 50,
587
+ "batch_size": 50,
588
+ "use_async_envs": false
589
+ },
590
+ "wandb": {
591
+ "enable": true,
592
+ "disable_artifact": false,
593
+ "project": "aic_sfp_dit_diffusion_embedding",
594
+ "entity": "gustave-charles-s-eth-z-rich",
595
+ "notes": null,
596
+ "run_id": "2ipoouae",
597
+ "mode": null,
598
+ "add_tags": true
599
+ },
600
+ "peft": null,
601
+ "use_rabc": false,
602
+ "rabc_progress_path": null,
603
+ "rabc_kappa": 0.01,
604
+ "rabc_epsilon": 1e-06,
605
+ "rabc_head_mode": "sparse",
606
+ "rename_map": {},
607
+ "checkpoint_path": null
608
+ }
checkpoints/010000/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,308 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 4.877641290737884e-05,
4
+ "betas": [
5
+ 0.95,
6
+ 0.999
7
+ ],
8
+ "eps": 1e-08,
9
+ "weight_decay": 0.0001,
10
+ "amsgrad": false,
11
+ "maximize": false,
12
+ "foreach": null,
13
+ "capturable": false,
14
+ "differentiable": false,
15
+ "fused": null,
16
+ "decoupled_weight_decay": false,
17
+ "initial_lr": 5e-05,
18
+ "params": [
19
+ 0,
20
+ 1,
21
+ 2,
22
+ 3,
23
+ 4,
24
+ 5,
25
+ 6,
26
+ 7,
27
+ 8,
28
+ 9,
29
+ 10,
30
+ 11,
31
+ 12,
32
+ 13,
33
+ 14,
34
+ 15,
35
+ 16,
36
+ 17,
37
+ 18,
38
+ 19,
39
+ 20,
40
+ 21,
41
+ 22,
42
+ 23,
43
+ 24,
44
+ 25,
45
+ 26,
46
+ 27,
47
+ 28,
48
+ 29,
49
+ 30,
50
+ 31,
51
+ 32,
52
+ 33,
53
+ 34,
54
+ 35,
55
+ 36,
56
+ 37,
57
+ 38,
58
+ 39,
59
+ 40,
60
+ 41,
61
+ 42,
62
+ 43,
63
+ 44,
64
+ 45,
65
+ 46,
66
+ 47,
67
+ 48,
68
+ 49,
69
+ 50,
70
+ 51,
71
+ 52,
72
+ 53,
73
+ 54,
74
+ 55,
75
+ 56,
76
+ 57,
77
+ 58,
78
+ 59,
79
+ 60,
80
+ 61,
81
+ 62,
82
+ 63,
83
+ 64,
84
+ 65,
85
+ 66,
86
+ 67,
87
+ 68
88
+ ]
89
+ },
90
+ {
91
+ "lr": 4.8776412907378845e-06,
92
+ "betas": [
93
+ 0.95,
94
+ 0.999
95
+ ],
96
+ "eps": 1e-08,
97
+ "weight_decay": 0.0001,
98
+ "amsgrad": false,
99
+ "maximize": false,
100
+ "foreach": null,
101
+ "capturable": false,
102
+ "differentiable": false,
103
+ "fused": null,
104
+ "decoupled_weight_decay": false,
105
+ "initial_lr": 5e-06,
106
+ "params": [
107
+ 69,
108
+ 70,
109
+ 71,
110
+ 72,
111
+ 73,
112
+ 74,
113
+ 75,
114
+ 76,
115
+ 77,
116
+ 78,
117
+ 79,
118
+ 80,
119
+ 81,
120
+ 82,
121
+ 83,
122
+ 84,
123
+ 85,
124
+ 86,
125
+ 87,
126
+ 88,
127
+ 89,
128
+ 90,
129
+ 91,
130
+ 92,
131
+ 93,
132
+ 94,
133
+ 95,
134
+ 96,
135
+ 97,
136
+ 98,
137
+ 99,
138
+ 100,
139
+ 101,
140
+ 102,
141
+ 103,
142
+ 104,
143
+ 105,
144
+ 106,
145
+ 107,
146
+ 108,
147
+ 109,
148
+ 110,
149
+ 111,
150
+ 112,
151
+ 113,
152
+ 114,
153
+ 115,
154
+ 116,
155
+ 117,
156
+ 118,
157
+ 119,
158
+ 120,
159
+ 121,
160
+ 122,
161
+ 123,
162
+ 124,
163
+ 125,
164
+ 126,
165
+ 127,
166
+ 128,
167
+ 129,
168
+ 130,
169
+ 131,
170
+ 132,
171
+ 133,
172
+ 134,
173
+ 135,
174
+ 136,
175
+ 137,
176
+ 138,
177
+ 139,
178
+ 140,
179
+ 141,
180
+ 142,
181
+ 143,
182
+ 144,
183
+ 145,
184
+ 146,
185
+ 147,
186
+ 148,
187
+ 149,
188
+ 150,
189
+ 151,
190
+ 152,
191
+ 153,
192
+ 154,
193
+ 155,
194
+ 156,
195
+ 157,
196
+ 158,
197
+ 159,
198
+ 160,
199
+ 161,
200
+ 162,
201
+ 163,
202
+ 164,
203
+ 165,
204
+ 166,
205
+ 167,
206
+ 168,
207
+ 169,
208
+ 170,
209
+ 171,
210
+ 172,
211
+ 173,
212
+ 174,
213
+ 175,
214
+ 176,
215
+ 177,
216
+ 178,
217
+ 179,
218
+ 180,
219
+ 181,
220
+ 182,
221
+ 183,
222
+ 184,
223
+ 185,
224
+ 186,
225
+ 187,
226
+ 188,
227
+ 189,
228
+ 190,
229
+ 191,
230
+ 192,
231
+ 193,
232
+ 194,
233
+ 195,
234
+ 196,
235
+ 197,
236
+ 198,
237
+ 199,
238
+ 200,
239
+ 201,
240
+ 202,
241
+ 203,
242
+ 204,
243
+ 205,
244
+ 206,
245
+ 207,
246
+ 208,
247
+ 209,
248
+ 210,
249
+ 211,
250
+ 212,
251
+ 213,
252
+ 214,
253
+ 215,
254
+ 216,
255
+ 217,
256
+ 218,
257
+ 219,
258
+ 220,
259
+ 221,
260
+ 222,
261
+ 223,
262
+ 224,
263
+ 225,
264
+ 226,
265
+ 227,
266
+ 228,
267
+ 229,
268
+ 230,
269
+ 231,
270
+ 232,
271
+ 233,
272
+ 234,
273
+ 235,
274
+ 236,
275
+ 237,
276
+ 238,
277
+ 239,
278
+ 240,
279
+ 241,
280
+ 242,
281
+ 243,
282
+ 244,
283
+ 245,
284
+ 246,
285
+ 247,
286
+ 248,
287
+ 249,
288
+ 250,
289
+ 251,
290
+ 252,
291
+ 253,
292
+ 254,
293
+ 255,
294
+ 256,
295
+ 257,
296
+ 258,
297
+ 259,
298
+ 260,
299
+ 261,
300
+ 262,
301
+ 263,
302
+ 264,
303
+ 265,
304
+ 266,
305
+ 267
306
+ ]
307
+ }
308
+ ]
checkpoints/010000/training_state/optimizer_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d3cc08359f9ed344ad5f25761fc7f30d00f89215ec58d439a5686e55c0a2f65c
3
+ size 1715911736
checkpoints/010000/training_state/rng_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d2cda57fb4963bfd45fa4f0dbf6f718ea9b4c8fc3164bd5d6086a7141bc3d939
3
+ size 15708
checkpoints/010000/training_state/scheduler_state.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_lrs": [
3
+ 5e-05,
4
+ 5e-06
5
+ ],
6
+ "last_epoch": 10000,
7
+ "_step_count": 10001,
8
+ "_get_lr_called_within_step": false,
9
+ "_last_lr": [
10
+ 4.877641290737884e-05,
11
+ 4.8776412907378845e-06
12
+ ],
13
+ "lr_lambdas": [
14
+ null,
15
+ null
16
+ ]
17
+ }
checkpoints/010000/training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 10000
3
+ }
checkpoints/020000/pretrained_model/config.json ADDED
@@ -0,0 +1,121 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "aic_multi_task_dit",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.images.left_camera": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 256,
10
+ 288
11
+ ]
12
+ },
13
+ "observation.images.center_camera": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 256,
18
+ 288
19
+ ]
20
+ },
21
+ "observation.images.right_camera": {
22
+ "type": "VISUAL",
23
+ "shape": [
24
+ 3,
25
+ 256,
26
+ 288
27
+ ]
28
+ },
29
+ "observation.state": {
30
+ "type": "STATE",
31
+ "shape": [
32
+ 26
33
+ ]
34
+ },
35
+ "observation.wrench": {
36
+ "type": "STATE",
37
+ "shape": [
38
+ 6
39
+ ]
40
+ }
41
+ },
42
+ "output_features": {
43
+ "action": {
44
+ "type": "ACTION",
45
+ "shape": [
46
+ 7
47
+ ]
48
+ }
49
+ },
50
+ "device": "cuda",
51
+ "use_amp": false,
52
+ "use_peft": false,
53
+ "push_to_hub": false,
54
+ "repo_id": null,
55
+ "private": null,
56
+ "tags": null,
57
+ "license": null,
58
+ "pretrained_path": null,
59
+ "horizon": 32,
60
+ "n_action_steps": 24,
61
+ "task_conditioning": "embedding",
62
+ "num_tasks": 10,
63
+ "objective": "diffusion",
64
+ "noise_scheduler_type": "DDPM",
65
+ "num_train_timesteps": 100,
66
+ "beta_schedule": "squaredcos_cap_v2",
67
+ "beta_start": 0.0001,
68
+ "beta_end": 0.02,
69
+ "prediction_type": "epsilon",
70
+ "clip_sample": true,
71
+ "clip_sample_range": 1.0,
72
+ "num_inference_steps": null,
73
+ "sigma_min": 0.0,
74
+ "num_integration_steps": 100,
75
+ "integration_method": "euler",
76
+ "timestep_sampling_strategy": "beta",
77
+ "timestep_sampling_s": 0.999,
78
+ "timestep_sampling_alpha": 1.5,
79
+ "timestep_sampling_beta": 1.0,
80
+ "hidden_dim": 512,
81
+ "num_layers": 6,
82
+ "num_heads": 8,
83
+ "dropout": 0.1,
84
+ "use_positional_encoding": false,
85
+ "timestep_embed_dim": 256,
86
+ "use_rope": true,
87
+ "rope_base": 10000.0,
88
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
89
+ "use_separate_rgb_encoder_per_camera": false,
90
+ "vision_encoder_lr_multiplier": 0.1,
91
+ "image_resize_shape": [
92
+ 232,
93
+ 232
94
+ ],
95
+ "image_crop_shape": [
96
+ 224,
97
+ 224
98
+ ],
99
+ "image_crop_is_random": true,
100
+ "text_encoder_name": "openai/clip-vit-base-patch16",
101
+ "tokenizer_max_length": 77,
102
+ "tokenizer_padding": "max_length",
103
+ "tokenizer_padding_side": "right",
104
+ "tokenizer_truncation": true,
105
+ "normalization_mapping": {
106
+ "VISUAL": "MEAN_STD",
107
+ "STATE": "MIN_MAX",
108
+ "ACTION": "MIN_MAX"
109
+ },
110
+ "optimizer_lr": 5e-05,
111
+ "optimizer_betas": [
112
+ 0.95,
113
+ 0.999
114
+ ],
115
+ "optimizer_eps": 1e-08,
116
+ "optimizer_weight_decay": 0.0001,
117
+ "scheduler_name": "cosine",
118
+ "scheduler_warmup_steps": 0,
119
+ "do_mask_loss_for_padding": false,
120
+ "drop_n_last_frames": 7
121
+ }
checkpoints/020000/pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5bbabad8cbd223b39a19d45f73c02495d3bb29ab950decf2d892ee9869673a1e
3
+ size 857967220
checkpoints/020000/pretrained_model/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": "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
+ }
checkpoints/020000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d58fa83446e4b0bdfc09a02565d49fca3dd894c73486df2d1ce3a22e752cd6
3
+ size 11368
checkpoints/020000/pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.left_camera": {
27
+ "type": "VISUAL",
28
+ "shape": [
29
+ 3,
30
+ 256,
31
+ 288
32
+ ]
33
+ },
34
+ "observation.images.center_camera": {
35
+ "type": "VISUAL",
36
+ "shape": [
37
+ 3,
38
+ 256,
39
+ 288
40
+ ]
41
+ },
42
+ "observation.images.right_camera": {
43
+ "type": "VISUAL",
44
+ "shape": [
45
+ 3,
46
+ 256,
47
+ 288
48
+ ]
49
+ },
50
+ "observation.state": {
51
+ "type": "STATE",
52
+ "shape": [
53
+ 26
54
+ ]
55
+ },
56
+ "observation.wrench": {
57
+ "type": "STATE",
58
+ "shape": [
59
+ 6
60
+ ]
61
+ },
62
+ "action": {
63
+ "type": "ACTION",
64
+ "shape": [
65
+ 7
66
+ ]
67
+ }
68
+ },
69
+ "norm_map": {
70
+ "VISUAL": "MEAN_STD",
71
+ "STATE": "MIN_MAX",
72
+ "ACTION": "MIN_MAX"
73
+ }
74
+ },
75
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
76
+ }
77
+ ]
78
+ }
checkpoints/020000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d58fa83446e4b0bdfc09a02565d49fca3dd894c73486df2d1ce3a22e752cd6
3
+ size 11368
checkpoints/020000/pretrained_model/train_config.json ADDED
@@ -0,0 +1,608 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "local/local_aic_cheatcode_demos_dit_snap",
4
+ "root": "/opt/dlami/nvme/lerobot/local_aic_cheatcode_demos_dit_snap",
5
+ "episodes": [
6
+ 3,
7
+ 5,
8
+ 8,
9
+ 157,
10
+ 183,
11
+ 184,
12
+ 185,
13
+ 186,
14
+ 187,
15
+ 188,
16
+ 189,
17
+ 190,
18
+ 192,
19
+ 193,
20
+ 194,
21
+ 195,
22
+ 196,
23
+ 198,
24
+ 199,
25
+ 200,
26
+ 201,
27
+ 203,
28
+ 204,
29
+ 205,
30
+ 206,
31
+ 207,
32
+ 208,
33
+ 209,
34
+ 210,
35
+ 211,
36
+ 213,
37
+ 214,
38
+ 215,
39
+ 216,
40
+ 217,
41
+ 218,
42
+ 219,
43
+ 220,
44
+ 221,
45
+ 222,
46
+ 223,
47
+ 224,
48
+ 225,
49
+ 227,
50
+ 228,
51
+ 229,
52
+ 230,
53
+ 232,
54
+ 233,
55
+ 234,
56
+ 236,
57
+ 238,
58
+ 239,
59
+ 241,
60
+ 242,
61
+ 243,
62
+ 244,
63
+ 245,
64
+ 246,
65
+ 247,
66
+ 248,
67
+ 249,
68
+ 250,
69
+ 251,
70
+ 252,
71
+ 253,
72
+ 255,
73
+ 256,
74
+ 257,
75
+ 258,
76
+ 259,
77
+ 260,
78
+ 262,
79
+ 263,
80
+ 264,
81
+ 265,
82
+ 266,
83
+ 267,
84
+ 268,
85
+ 269,
86
+ 270,
87
+ 271,
88
+ 272,
89
+ 273,
90
+ 274,
91
+ 276,
92
+ 277,
93
+ 278,
94
+ 279,
95
+ 280,
96
+ 281,
97
+ 283,
98
+ 284,
99
+ 285,
100
+ 286,
101
+ 287,
102
+ 288,
103
+ 289,
104
+ 290,
105
+ 291,
106
+ 292,
107
+ 293,
108
+ 294,
109
+ 295,
110
+ 298,
111
+ 299,
112
+ 300,
113
+ 301,
114
+ 302,
115
+ 303,
116
+ 304,
117
+ 305,
118
+ 306,
119
+ 307,
120
+ 308,
121
+ 309,
122
+ 310,
123
+ 311,
124
+ 312,
125
+ 313,
126
+ 314,
127
+ 315,
128
+ 316,
129
+ 317,
130
+ 318,
131
+ 319,
132
+ 321,
133
+ 322,
134
+ 323,
135
+ 325,
136
+ 327,
137
+ 328,
138
+ 329,
139
+ 330,
140
+ 332,
141
+ 333,
142
+ 334,
143
+ 335,
144
+ 336,
145
+ 337,
146
+ 338,
147
+ 340,
148
+ 341,
149
+ 342,
150
+ 343,
151
+ 344,
152
+ 345,
153
+ 346,
154
+ 347,
155
+ 348,
156
+ 350,
157
+ 351,
158
+ 352,
159
+ 354,
160
+ 355,
161
+ 356,
162
+ 358,
163
+ 359,
164
+ 360,
165
+ 361,
166
+ 362,
167
+ 363,
168
+ 364,
169
+ 366,
170
+ 367,
171
+ 368,
172
+ 369,
173
+ 370,
174
+ 371,
175
+ 372,
176
+ 373,
177
+ 374,
178
+ 375,
179
+ 377,
180
+ 378,
181
+ 379,
182
+ 380,
183
+ 381,
184
+ 384,
185
+ 385,
186
+ 386,
187
+ 387,
188
+ 388,
189
+ 389,
190
+ 390,
191
+ 391,
192
+ 393,
193
+ 395,
194
+ 397,
195
+ 399,
196
+ 400,
197
+ 401,
198
+ 402,
199
+ 404,
200
+ 405,
201
+ 407,
202
+ 408,
203
+ 409,
204
+ 410,
205
+ 411,
206
+ 413,
207
+ 414,
208
+ 415,
209
+ 416,
210
+ 417,
211
+ 419,
212
+ 420,
213
+ 421,
214
+ 422,
215
+ 423,
216
+ 425,
217
+ 430,
218
+ 431,
219
+ 433,
220
+ 434,
221
+ 436,
222
+ 437,
223
+ 439,
224
+ 440,
225
+ 442,
226
+ 444,
227
+ 447,
228
+ 449,
229
+ 450,
230
+ 452,
231
+ 453,
232
+ 454,
233
+ 455,
234
+ 457,
235
+ 458,
236
+ 460,
237
+ 462,
238
+ 463,
239
+ 464,
240
+ 465,
241
+ 466,
242
+ 467,
243
+ 470,
244
+ 471,
245
+ 473,
246
+ 474,
247
+ 475,
248
+ 476,
249
+ 478,
250
+ 479,
251
+ 480,
252
+ 482,
253
+ 485,
254
+ 490,
255
+ 491,
256
+ 492,
257
+ 497,
258
+ 498,
259
+ 500,
260
+ 501,
261
+ 502,
262
+ 504,
263
+ 507,
264
+ 508,
265
+ 509,
266
+ 512,
267
+ 514,
268
+ 515,
269
+ 516,
270
+ 517,
271
+ 518,
272
+ 520,
273
+ 522,
274
+ 524,
275
+ 527,
276
+ 529,
277
+ 530,
278
+ 532,
279
+ 533,
280
+ 534,
281
+ 535,
282
+ 537,
283
+ 538,
284
+ 539,
285
+ 540,
286
+ 541,
287
+ 542,
288
+ 543,
289
+ 544,
290
+ 546,
291
+ 548,
292
+ 549,
293
+ 550,
294
+ 551,
295
+ 553,
296
+ 555,
297
+ 556,
298
+ 557,
299
+ 558,
300
+ 559,
301
+ 560,
302
+ 562,
303
+ 563,
304
+ 565,
305
+ 566,
306
+ 569,
307
+ 570,
308
+ 571,
309
+ 572,
310
+ 575,
311
+ 576,
312
+ 577,
313
+ 578,
314
+ 580,
315
+ 581,
316
+ 582,
317
+ 584,
318
+ 585,
319
+ 586,
320
+ 587,
321
+ 588,
322
+ 589,
323
+ 590,
324
+ 591,
325
+ 593,
326
+ 594,
327
+ 595,
328
+ 596,
329
+ 598,
330
+ 600,
331
+ 601,
332
+ 602,
333
+ 603,
334
+ 606,
335
+ 607,
336
+ 608,
337
+ 609,
338
+ 612,
339
+ 613,
340
+ 615,
341
+ 616,
342
+ 617,
343
+ 619,
344
+ 620,
345
+ 621,
346
+ 623,
347
+ 624,
348
+ 627,
349
+ 628,
350
+ 631,
351
+ 632,
352
+ 633,
353
+ 634,
354
+ 635,
355
+ 636
356
+ ],
357
+ "image_transforms": {
358
+ "enable": true,
359
+ "max_num_transforms": 3,
360
+ "random_order": true,
361
+ "tfs": {
362
+ "brightness": {
363
+ "weight": 1.0,
364
+ "type": "ColorJitter",
365
+ "kwargs": {
366
+ "brightness": [
367
+ 0.9,
368
+ 1.1
369
+ ]
370
+ }
371
+ },
372
+ "contrast": {
373
+ "weight": 1.0,
374
+ "type": "ColorJitter",
375
+ "kwargs": {
376
+ "contrast": [
377
+ 0.9,
378
+ 1.1
379
+ ]
380
+ }
381
+ },
382
+ "saturation": {
383
+ "weight": 1.0,
384
+ "type": "ColorJitter",
385
+ "kwargs": {
386
+ "saturation": [
387
+ 0.95,
388
+ 1.05
389
+ ]
390
+ }
391
+ },
392
+ "hue": {
393
+ "weight": 1.0,
394
+ "type": "ColorJitter",
395
+ "kwargs": {
396
+ "hue": [
397
+ -0.02,
398
+ 0.02
399
+ ]
400
+ }
401
+ },
402
+ "sharpness": {
403
+ "weight": 1.0,
404
+ "type": "SharpnessJitter",
405
+ "kwargs": {
406
+ "sharpness": [
407
+ 0.9,
408
+ 1.1
409
+ ]
410
+ }
411
+ },
412
+ "affine": {
413
+ "weight": 0.0,
414
+ "type": "RandomAffine",
415
+ "kwargs": {
416
+ "degrees": [
417
+ -5.0,
418
+ 5.0
419
+ ],
420
+ "translate": [
421
+ 0.05,
422
+ 0.05
423
+ ]
424
+ }
425
+ }
426
+ }
427
+ },
428
+ "revision": null,
429
+ "use_imagenet_stats": true,
430
+ "video_backend": "pyav",
431
+ "streaming": false
432
+ },
433
+ "env": null,
434
+ "policy": {
435
+ "type": "aic_multi_task_dit",
436
+ "n_obs_steps": 2,
437
+ "input_features": {
438
+ "observation.images.left_camera": {
439
+ "type": "VISUAL",
440
+ "shape": [
441
+ 3,
442
+ 256,
443
+ 288
444
+ ]
445
+ },
446
+ "observation.images.center_camera": {
447
+ "type": "VISUAL",
448
+ "shape": [
449
+ 3,
450
+ 256,
451
+ 288
452
+ ]
453
+ },
454
+ "observation.images.right_camera": {
455
+ "type": "VISUAL",
456
+ "shape": [
457
+ 3,
458
+ 256,
459
+ 288
460
+ ]
461
+ },
462
+ "observation.state": {
463
+ "type": "STATE",
464
+ "shape": [
465
+ 26
466
+ ]
467
+ },
468
+ "observation.wrench": {
469
+ "type": "STATE",
470
+ "shape": [
471
+ 6
472
+ ]
473
+ }
474
+ },
475
+ "output_features": {
476
+ "action": {
477
+ "type": "ACTION",
478
+ "shape": [
479
+ 7
480
+ ]
481
+ }
482
+ },
483
+ "device": "cuda",
484
+ "use_amp": false,
485
+ "use_peft": false,
486
+ "push_to_hub": false,
487
+ "repo_id": null,
488
+ "private": null,
489
+ "tags": null,
490
+ "license": null,
491
+ "pretrained_path": null,
492
+ "horizon": 32,
493
+ "n_action_steps": 24,
494
+ "task_conditioning": "embedding",
495
+ "num_tasks": 10,
496
+ "objective": "diffusion",
497
+ "noise_scheduler_type": "DDPM",
498
+ "num_train_timesteps": 100,
499
+ "beta_schedule": "squaredcos_cap_v2",
500
+ "beta_start": 0.0001,
501
+ "beta_end": 0.02,
502
+ "prediction_type": "epsilon",
503
+ "clip_sample": true,
504
+ "clip_sample_range": 1.0,
505
+ "num_inference_steps": null,
506
+ "sigma_min": 0.0,
507
+ "num_integration_steps": 100,
508
+ "integration_method": "euler",
509
+ "timestep_sampling_strategy": "beta",
510
+ "timestep_sampling_s": 0.999,
511
+ "timestep_sampling_alpha": 1.5,
512
+ "timestep_sampling_beta": 1.0,
513
+ "hidden_dim": 512,
514
+ "num_layers": 6,
515
+ "num_heads": 8,
516
+ "dropout": 0.1,
517
+ "use_positional_encoding": false,
518
+ "timestep_embed_dim": 256,
519
+ "use_rope": true,
520
+ "rope_base": 10000.0,
521
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
522
+ "use_separate_rgb_encoder_per_camera": false,
523
+ "vision_encoder_lr_multiplier": 0.1,
524
+ "image_resize_shape": [
525
+ 232,
526
+ 232
527
+ ],
528
+ "image_crop_shape": [
529
+ 224,
530
+ 224
531
+ ],
532
+ "image_crop_is_random": true,
533
+ "text_encoder_name": "openai/clip-vit-base-patch16",
534
+ "tokenizer_max_length": 77,
535
+ "tokenizer_padding": "max_length",
536
+ "tokenizer_padding_side": "right",
537
+ "tokenizer_truncation": true,
538
+ "normalization_mapping": {
539
+ "VISUAL": "MEAN_STD",
540
+ "STATE": "MIN_MAX",
541
+ "ACTION": "MIN_MAX"
542
+ },
543
+ "optimizer_lr": 5e-05,
544
+ "optimizer_betas": [
545
+ 0.95,
546
+ 0.999
547
+ ],
548
+ "optimizer_eps": 1e-08,
549
+ "optimizer_weight_decay": 0.0001,
550
+ "scheduler_name": "cosine",
551
+ "scheduler_warmup_steps": 0,
552
+ "do_mask_loss_for_padding": false,
553
+ "drop_n_last_frames": 7
554
+ },
555
+ "output_dir": "/opt/dlami/nvme/outputs/dit_diffusion_20260511_105440",
556
+ "job_name": "aic_multi_task_dit",
557
+ "resume": false,
558
+ "seed": 42,
559
+ "cudnn_deterministic": false,
560
+ "num_workers": 4,
561
+ "batch_size": 12,
562
+ "steps": 100000,
563
+ "eval_freq": 0,
564
+ "log_freq": 100,
565
+ "tolerance_s": 0.1,
566
+ "save_checkpoint": true,
567
+ "save_freq": 10000,
568
+ "use_policy_training_preset": true,
569
+ "optimizer": {
570
+ "type": "adam",
571
+ "lr": 5e-05,
572
+ "weight_decay": 0.0001,
573
+ "grad_clip_norm": 10.0,
574
+ "betas": [
575
+ 0.95,
576
+ 0.999
577
+ ],
578
+ "eps": 1e-08
579
+ },
580
+ "scheduler": {
581
+ "type": "diffuser",
582
+ "num_warmup_steps": 0,
583
+ "name": "cosine"
584
+ },
585
+ "eval": {
586
+ "n_episodes": 50,
587
+ "batch_size": 50,
588
+ "use_async_envs": false
589
+ },
590
+ "wandb": {
591
+ "enable": true,
592
+ "disable_artifact": false,
593
+ "project": "aic_sfp_dit_diffusion_embedding",
594
+ "entity": "gustave-charles-s-eth-z-rich",
595
+ "notes": null,
596
+ "run_id": "2ipoouae",
597
+ "mode": null,
598
+ "add_tags": true
599
+ },
600
+ "peft": null,
601
+ "use_rabc": false,
602
+ "rabc_progress_path": null,
603
+ "rabc_kappa": 0.01,
604
+ "rabc_epsilon": 1e-06,
605
+ "rabc_head_mode": "sparse",
606
+ "rename_map": {},
607
+ "checkpoint_path": null
608
+ }
checkpoints/020000/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,308 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 4.522542485937369e-05,
4
+ "betas": [
5
+ 0.95,
6
+ 0.999
7
+ ],
8
+ "eps": 1e-08,
9
+ "weight_decay": 0.0001,
10
+ "amsgrad": false,
11
+ "maximize": false,
12
+ "foreach": null,
13
+ "capturable": false,
14
+ "differentiable": false,
15
+ "fused": null,
16
+ "decoupled_weight_decay": false,
17
+ "initial_lr": 5e-05,
18
+ "params": [
19
+ 0,
20
+ 1,
21
+ 2,
22
+ 3,
23
+ 4,
24
+ 5,
25
+ 6,
26
+ 7,
27
+ 8,
28
+ 9,
29
+ 10,
30
+ 11,
31
+ 12,
32
+ 13,
33
+ 14,
34
+ 15,
35
+ 16,
36
+ 17,
37
+ 18,
38
+ 19,
39
+ 20,
40
+ 21,
41
+ 22,
42
+ 23,
43
+ 24,
44
+ 25,
45
+ 26,
46
+ 27,
47
+ 28,
48
+ 29,
49
+ 30,
50
+ 31,
51
+ 32,
52
+ 33,
53
+ 34,
54
+ 35,
55
+ 36,
56
+ 37,
57
+ 38,
58
+ 39,
59
+ 40,
60
+ 41,
61
+ 42,
62
+ 43,
63
+ 44,
64
+ 45,
65
+ 46,
66
+ 47,
67
+ 48,
68
+ 49,
69
+ 50,
70
+ 51,
71
+ 52,
72
+ 53,
73
+ 54,
74
+ 55,
75
+ 56,
76
+ 57,
77
+ 58,
78
+ 59,
79
+ 60,
80
+ 61,
81
+ 62,
82
+ 63,
83
+ 64,
84
+ 65,
85
+ 66,
86
+ 67,
87
+ 68
88
+ ]
89
+ },
90
+ {
91
+ "lr": 4.522542485937369e-06,
92
+ "betas": [
93
+ 0.95,
94
+ 0.999
95
+ ],
96
+ "eps": 1e-08,
97
+ "weight_decay": 0.0001,
98
+ "amsgrad": false,
99
+ "maximize": false,
100
+ "foreach": null,
101
+ "capturable": false,
102
+ "differentiable": false,
103
+ "fused": null,
104
+ "decoupled_weight_decay": false,
105
+ "initial_lr": 5e-06,
106
+ "params": [
107
+ 69,
108
+ 70,
109
+ 71,
110
+ 72,
111
+ 73,
112
+ 74,
113
+ 75,
114
+ 76,
115
+ 77,
116
+ 78,
117
+ 79,
118
+ 80,
119
+ 81,
120
+ 82,
121
+ 83,
122
+ 84,
123
+ 85,
124
+ 86,
125
+ 87,
126
+ 88,
127
+ 89,
128
+ 90,
129
+ 91,
130
+ 92,
131
+ 93,
132
+ 94,
133
+ 95,
134
+ 96,
135
+ 97,
136
+ 98,
137
+ 99,
138
+ 100,
139
+ 101,
140
+ 102,
141
+ 103,
142
+ 104,
143
+ 105,
144
+ 106,
145
+ 107,
146
+ 108,
147
+ 109,
148
+ 110,
149
+ 111,
150
+ 112,
151
+ 113,
152
+ 114,
153
+ 115,
154
+ 116,
155
+ 117,
156
+ 118,
157
+ 119,
158
+ 120,
159
+ 121,
160
+ 122,
161
+ 123,
162
+ 124,
163
+ 125,
164
+ 126,
165
+ 127,
166
+ 128,
167
+ 129,
168
+ 130,
169
+ 131,
170
+ 132,
171
+ 133,
172
+ 134,
173
+ 135,
174
+ 136,
175
+ 137,
176
+ 138,
177
+ 139,
178
+ 140,
179
+ 141,
180
+ 142,
181
+ 143,
182
+ 144,
183
+ 145,
184
+ 146,
185
+ 147,
186
+ 148,
187
+ 149,
188
+ 150,
189
+ 151,
190
+ 152,
191
+ 153,
192
+ 154,
193
+ 155,
194
+ 156,
195
+ 157,
196
+ 158,
197
+ 159,
198
+ 160,
199
+ 161,
200
+ 162,
201
+ 163,
202
+ 164,
203
+ 165,
204
+ 166,
205
+ 167,
206
+ 168,
207
+ 169,
208
+ 170,
209
+ 171,
210
+ 172,
211
+ 173,
212
+ 174,
213
+ 175,
214
+ 176,
215
+ 177,
216
+ 178,
217
+ 179,
218
+ 180,
219
+ 181,
220
+ 182,
221
+ 183,
222
+ 184,
223
+ 185,
224
+ 186,
225
+ 187,
226
+ 188,
227
+ 189,
228
+ 190,
229
+ 191,
230
+ 192,
231
+ 193,
232
+ 194,
233
+ 195,
234
+ 196,
235
+ 197,
236
+ 198,
237
+ 199,
238
+ 200,
239
+ 201,
240
+ 202,
241
+ 203,
242
+ 204,
243
+ 205,
244
+ 206,
245
+ 207,
246
+ 208,
247
+ 209,
248
+ 210,
249
+ 211,
250
+ 212,
251
+ 213,
252
+ 214,
253
+ 215,
254
+ 216,
255
+ 217,
256
+ 218,
257
+ 219,
258
+ 220,
259
+ 221,
260
+ 222,
261
+ 223,
262
+ 224,
263
+ 225,
264
+ 226,
265
+ 227,
266
+ 228,
267
+ 229,
268
+ 230,
269
+ 231,
270
+ 232,
271
+ 233,
272
+ 234,
273
+ 235,
274
+ 236,
275
+ 237,
276
+ 238,
277
+ 239,
278
+ 240,
279
+ 241,
280
+ 242,
281
+ 243,
282
+ 244,
283
+ 245,
284
+ 246,
285
+ 247,
286
+ 248,
287
+ 249,
288
+ 250,
289
+ 251,
290
+ 252,
291
+ 253,
292
+ 254,
293
+ 255,
294
+ 256,
295
+ 257,
296
+ 258,
297
+ 259,
298
+ 260,
299
+ 261,
300
+ 262,
301
+ 263,
302
+ 264,
303
+ 265,
304
+ 266,
305
+ 267
306
+ ]
307
+ }
308
+ ]
checkpoints/020000/training_state/optimizer_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ed26cf37ed67dc91b0c0c4c085eb9f42d052191b5935c1ac8a74adb5791aaab5
3
+ size 1715911736
checkpoints/020000/training_state/rng_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f1fd30d6d274d117d510e9a89dfabe0c19405650c88173a5b5ff37fc1555354
3
+ size 15708
checkpoints/020000/training_state/scheduler_state.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_lrs": [
3
+ 5e-05,
4
+ 5e-06
5
+ ],
6
+ "last_epoch": 20000,
7
+ "_step_count": 20001,
8
+ "_get_lr_called_within_step": false,
9
+ "_last_lr": [
10
+ 4.522542485937369e-05,
11
+ 4.522542485937369e-06
12
+ ],
13
+ "lr_lambdas": [
14
+ null,
15
+ null
16
+ ]
17
+ }
checkpoints/020000/training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 20000
3
+ }
checkpoints/030000/pretrained_model/config.json ADDED
@@ -0,0 +1,121 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "aic_multi_task_dit",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.images.left_camera": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 256,
10
+ 288
11
+ ]
12
+ },
13
+ "observation.images.center_camera": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 256,
18
+ 288
19
+ ]
20
+ },
21
+ "observation.images.right_camera": {
22
+ "type": "VISUAL",
23
+ "shape": [
24
+ 3,
25
+ 256,
26
+ 288
27
+ ]
28
+ },
29
+ "observation.state": {
30
+ "type": "STATE",
31
+ "shape": [
32
+ 26
33
+ ]
34
+ },
35
+ "observation.wrench": {
36
+ "type": "STATE",
37
+ "shape": [
38
+ 6
39
+ ]
40
+ }
41
+ },
42
+ "output_features": {
43
+ "action": {
44
+ "type": "ACTION",
45
+ "shape": [
46
+ 7
47
+ ]
48
+ }
49
+ },
50
+ "device": "cuda",
51
+ "use_amp": false,
52
+ "use_peft": false,
53
+ "push_to_hub": false,
54
+ "repo_id": null,
55
+ "private": null,
56
+ "tags": null,
57
+ "license": null,
58
+ "pretrained_path": null,
59
+ "horizon": 32,
60
+ "n_action_steps": 24,
61
+ "task_conditioning": "embedding",
62
+ "num_tasks": 10,
63
+ "objective": "diffusion",
64
+ "noise_scheduler_type": "DDPM",
65
+ "num_train_timesteps": 100,
66
+ "beta_schedule": "squaredcos_cap_v2",
67
+ "beta_start": 0.0001,
68
+ "beta_end": 0.02,
69
+ "prediction_type": "epsilon",
70
+ "clip_sample": true,
71
+ "clip_sample_range": 1.0,
72
+ "num_inference_steps": null,
73
+ "sigma_min": 0.0,
74
+ "num_integration_steps": 100,
75
+ "integration_method": "euler",
76
+ "timestep_sampling_strategy": "beta",
77
+ "timestep_sampling_s": 0.999,
78
+ "timestep_sampling_alpha": 1.5,
79
+ "timestep_sampling_beta": 1.0,
80
+ "hidden_dim": 512,
81
+ "num_layers": 6,
82
+ "num_heads": 8,
83
+ "dropout": 0.1,
84
+ "use_positional_encoding": false,
85
+ "timestep_embed_dim": 256,
86
+ "use_rope": true,
87
+ "rope_base": 10000.0,
88
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
89
+ "use_separate_rgb_encoder_per_camera": false,
90
+ "vision_encoder_lr_multiplier": 0.1,
91
+ "image_resize_shape": [
92
+ 232,
93
+ 232
94
+ ],
95
+ "image_crop_shape": [
96
+ 224,
97
+ 224
98
+ ],
99
+ "image_crop_is_random": true,
100
+ "text_encoder_name": "openai/clip-vit-base-patch16",
101
+ "tokenizer_max_length": 77,
102
+ "tokenizer_padding": "max_length",
103
+ "tokenizer_padding_side": "right",
104
+ "tokenizer_truncation": true,
105
+ "normalization_mapping": {
106
+ "VISUAL": "MEAN_STD",
107
+ "STATE": "MIN_MAX",
108
+ "ACTION": "MIN_MAX"
109
+ },
110
+ "optimizer_lr": 5e-05,
111
+ "optimizer_betas": [
112
+ 0.95,
113
+ 0.999
114
+ ],
115
+ "optimizer_eps": 1e-08,
116
+ "optimizer_weight_decay": 0.0001,
117
+ "scheduler_name": "cosine",
118
+ "scheduler_warmup_steps": 0,
119
+ "do_mask_loss_for_padding": false,
120
+ "drop_n_last_frames": 7
121
+ }
checkpoints/030000/pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b13b176e48e3052ea1ffb9840d84ba9a8e5464171cae49ffeceb388e6e999088
3
+ size 857967220
checkpoints/030000/pretrained_model/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": "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
+ }
checkpoints/030000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d58fa83446e4b0bdfc09a02565d49fca3dd894c73486df2d1ce3a22e752cd6
3
+ size 11368
checkpoints/030000/pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.left_camera": {
27
+ "type": "VISUAL",
28
+ "shape": [
29
+ 3,
30
+ 256,
31
+ 288
32
+ ]
33
+ },
34
+ "observation.images.center_camera": {
35
+ "type": "VISUAL",
36
+ "shape": [
37
+ 3,
38
+ 256,
39
+ 288
40
+ ]
41
+ },
42
+ "observation.images.right_camera": {
43
+ "type": "VISUAL",
44
+ "shape": [
45
+ 3,
46
+ 256,
47
+ 288
48
+ ]
49
+ },
50
+ "observation.state": {
51
+ "type": "STATE",
52
+ "shape": [
53
+ 26
54
+ ]
55
+ },
56
+ "observation.wrench": {
57
+ "type": "STATE",
58
+ "shape": [
59
+ 6
60
+ ]
61
+ },
62
+ "action": {
63
+ "type": "ACTION",
64
+ "shape": [
65
+ 7
66
+ ]
67
+ }
68
+ },
69
+ "norm_map": {
70
+ "VISUAL": "MEAN_STD",
71
+ "STATE": "MIN_MAX",
72
+ "ACTION": "MIN_MAX"
73
+ }
74
+ },
75
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
76
+ }
77
+ ]
78
+ }
checkpoints/030000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d58fa83446e4b0bdfc09a02565d49fca3dd894c73486df2d1ce3a22e752cd6
3
+ size 11368
checkpoints/030000/pretrained_model/train_config.json ADDED
@@ -0,0 +1,608 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "local/local_aic_cheatcode_demos_dit_snap",
4
+ "root": "/opt/dlami/nvme/lerobot/local_aic_cheatcode_demos_dit_snap",
5
+ "episodes": [
6
+ 3,
7
+ 5,
8
+ 8,
9
+ 157,
10
+ 183,
11
+ 184,
12
+ 185,
13
+ 186,
14
+ 187,
15
+ 188,
16
+ 189,
17
+ 190,
18
+ 192,
19
+ 193,
20
+ 194,
21
+ 195,
22
+ 196,
23
+ 198,
24
+ 199,
25
+ 200,
26
+ 201,
27
+ 203,
28
+ 204,
29
+ 205,
30
+ 206,
31
+ 207,
32
+ 208,
33
+ 209,
34
+ 210,
35
+ 211,
36
+ 213,
37
+ 214,
38
+ 215,
39
+ 216,
40
+ 217,
41
+ 218,
42
+ 219,
43
+ 220,
44
+ 221,
45
+ 222,
46
+ 223,
47
+ 224,
48
+ 225,
49
+ 227,
50
+ 228,
51
+ 229,
52
+ 230,
53
+ 232,
54
+ 233,
55
+ 234,
56
+ 236,
57
+ 238,
58
+ 239,
59
+ 241,
60
+ 242,
61
+ 243,
62
+ 244,
63
+ 245,
64
+ 246,
65
+ 247,
66
+ 248,
67
+ 249,
68
+ 250,
69
+ 251,
70
+ 252,
71
+ 253,
72
+ 255,
73
+ 256,
74
+ 257,
75
+ 258,
76
+ 259,
77
+ 260,
78
+ 262,
79
+ 263,
80
+ 264,
81
+ 265,
82
+ 266,
83
+ 267,
84
+ 268,
85
+ 269,
86
+ 270,
87
+ 271,
88
+ 272,
89
+ 273,
90
+ 274,
91
+ 276,
92
+ 277,
93
+ 278,
94
+ 279,
95
+ 280,
96
+ 281,
97
+ 283,
98
+ 284,
99
+ 285,
100
+ 286,
101
+ 287,
102
+ 288,
103
+ 289,
104
+ 290,
105
+ 291,
106
+ 292,
107
+ 293,
108
+ 294,
109
+ 295,
110
+ 298,
111
+ 299,
112
+ 300,
113
+ 301,
114
+ 302,
115
+ 303,
116
+ 304,
117
+ 305,
118
+ 306,
119
+ 307,
120
+ 308,
121
+ 309,
122
+ 310,
123
+ 311,
124
+ 312,
125
+ 313,
126
+ 314,
127
+ 315,
128
+ 316,
129
+ 317,
130
+ 318,
131
+ 319,
132
+ 321,
133
+ 322,
134
+ 323,
135
+ 325,
136
+ 327,
137
+ 328,
138
+ 329,
139
+ 330,
140
+ 332,
141
+ 333,
142
+ 334,
143
+ 335,
144
+ 336,
145
+ 337,
146
+ 338,
147
+ 340,
148
+ 341,
149
+ 342,
150
+ 343,
151
+ 344,
152
+ 345,
153
+ 346,
154
+ 347,
155
+ 348,
156
+ 350,
157
+ 351,
158
+ 352,
159
+ 354,
160
+ 355,
161
+ 356,
162
+ 358,
163
+ 359,
164
+ 360,
165
+ 361,
166
+ 362,
167
+ 363,
168
+ 364,
169
+ 366,
170
+ 367,
171
+ 368,
172
+ 369,
173
+ 370,
174
+ 371,
175
+ 372,
176
+ 373,
177
+ 374,
178
+ 375,
179
+ 377,
180
+ 378,
181
+ 379,
182
+ 380,
183
+ 381,
184
+ 384,
185
+ 385,
186
+ 386,
187
+ 387,
188
+ 388,
189
+ 389,
190
+ 390,
191
+ 391,
192
+ 393,
193
+ 395,
194
+ 397,
195
+ 399,
196
+ 400,
197
+ 401,
198
+ 402,
199
+ 404,
200
+ 405,
201
+ 407,
202
+ 408,
203
+ 409,
204
+ 410,
205
+ 411,
206
+ 413,
207
+ 414,
208
+ 415,
209
+ 416,
210
+ 417,
211
+ 419,
212
+ 420,
213
+ 421,
214
+ 422,
215
+ 423,
216
+ 425,
217
+ 430,
218
+ 431,
219
+ 433,
220
+ 434,
221
+ 436,
222
+ 437,
223
+ 439,
224
+ 440,
225
+ 442,
226
+ 444,
227
+ 447,
228
+ 449,
229
+ 450,
230
+ 452,
231
+ 453,
232
+ 454,
233
+ 455,
234
+ 457,
235
+ 458,
236
+ 460,
237
+ 462,
238
+ 463,
239
+ 464,
240
+ 465,
241
+ 466,
242
+ 467,
243
+ 470,
244
+ 471,
245
+ 473,
246
+ 474,
247
+ 475,
248
+ 476,
249
+ 478,
250
+ 479,
251
+ 480,
252
+ 482,
253
+ 485,
254
+ 490,
255
+ 491,
256
+ 492,
257
+ 497,
258
+ 498,
259
+ 500,
260
+ 501,
261
+ 502,
262
+ 504,
263
+ 507,
264
+ 508,
265
+ 509,
266
+ 512,
267
+ 514,
268
+ 515,
269
+ 516,
270
+ 517,
271
+ 518,
272
+ 520,
273
+ 522,
274
+ 524,
275
+ 527,
276
+ 529,
277
+ 530,
278
+ 532,
279
+ 533,
280
+ 534,
281
+ 535,
282
+ 537,
283
+ 538,
284
+ 539,
285
+ 540,
286
+ 541,
287
+ 542,
288
+ 543,
289
+ 544,
290
+ 546,
291
+ 548,
292
+ 549,
293
+ 550,
294
+ 551,
295
+ 553,
296
+ 555,
297
+ 556,
298
+ 557,
299
+ 558,
300
+ 559,
301
+ 560,
302
+ 562,
303
+ 563,
304
+ 565,
305
+ 566,
306
+ 569,
307
+ 570,
308
+ 571,
309
+ 572,
310
+ 575,
311
+ 576,
312
+ 577,
313
+ 578,
314
+ 580,
315
+ 581,
316
+ 582,
317
+ 584,
318
+ 585,
319
+ 586,
320
+ 587,
321
+ 588,
322
+ 589,
323
+ 590,
324
+ 591,
325
+ 593,
326
+ 594,
327
+ 595,
328
+ 596,
329
+ 598,
330
+ 600,
331
+ 601,
332
+ 602,
333
+ 603,
334
+ 606,
335
+ 607,
336
+ 608,
337
+ 609,
338
+ 612,
339
+ 613,
340
+ 615,
341
+ 616,
342
+ 617,
343
+ 619,
344
+ 620,
345
+ 621,
346
+ 623,
347
+ 624,
348
+ 627,
349
+ 628,
350
+ 631,
351
+ 632,
352
+ 633,
353
+ 634,
354
+ 635,
355
+ 636
356
+ ],
357
+ "image_transforms": {
358
+ "enable": true,
359
+ "max_num_transforms": 3,
360
+ "random_order": true,
361
+ "tfs": {
362
+ "brightness": {
363
+ "weight": 1.0,
364
+ "type": "ColorJitter",
365
+ "kwargs": {
366
+ "brightness": [
367
+ 0.9,
368
+ 1.1
369
+ ]
370
+ }
371
+ },
372
+ "contrast": {
373
+ "weight": 1.0,
374
+ "type": "ColorJitter",
375
+ "kwargs": {
376
+ "contrast": [
377
+ 0.9,
378
+ 1.1
379
+ ]
380
+ }
381
+ },
382
+ "saturation": {
383
+ "weight": 1.0,
384
+ "type": "ColorJitter",
385
+ "kwargs": {
386
+ "saturation": [
387
+ 0.95,
388
+ 1.05
389
+ ]
390
+ }
391
+ },
392
+ "hue": {
393
+ "weight": 1.0,
394
+ "type": "ColorJitter",
395
+ "kwargs": {
396
+ "hue": [
397
+ -0.02,
398
+ 0.02
399
+ ]
400
+ }
401
+ },
402
+ "sharpness": {
403
+ "weight": 1.0,
404
+ "type": "SharpnessJitter",
405
+ "kwargs": {
406
+ "sharpness": [
407
+ 0.9,
408
+ 1.1
409
+ ]
410
+ }
411
+ },
412
+ "affine": {
413
+ "weight": 0.0,
414
+ "type": "RandomAffine",
415
+ "kwargs": {
416
+ "degrees": [
417
+ -5.0,
418
+ 5.0
419
+ ],
420
+ "translate": [
421
+ 0.05,
422
+ 0.05
423
+ ]
424
+ }
425
+ }
426
+ }
427
+ },
428
+ "revision": null,
429
+ "use_imagenet_stats": true,
430
+ "video_backend": "pyav",
431
+ "streaming": false
432
+ },
433
+ "env": null,
434
+ "policy": {
435
+ "type": "aic_multi_task_dit",
436
+ "n_obs_steps": 2,
437
+ "input_features": {
438
+ "observation.images.left_camera": {
439
+ "type": "VISUAL",
440
+ "shape": [
441
+ 3,
442
+ 256,
443
+ 288
444
+ ]
445
+ },
446
+ "observation.images.center_camera": {
447
+ "type": "VISUAL",
448
+ "shape": [
449
+ 3,
450
+ 256,
451
+ 288
452
+ ]
453
+ },
454
+ "observation.images.right_camera": {
455
+ "type": "VISUAL",
456
+ "shape": [
457
+ 3,
458
+ 256,
459
+ 288
460
+ ]
461
+ },
462
+ "observation.state": {
463
+ "type": "STATE",
464
+ "shape": [
465
+ 26
466
+ ]
467
+ },
468
+ "observation.wrench": {
469
+ "type": "STATE",
470
+ "shape": [
471
+ 6
472
+ ]
473
+ }
474
+ },
475
+ "output_features": {
476
+ "action": {
477
+ "type": "ACTION",
478
+ "shape": [
479
+ 7
480
+ ]
481
+ }
482
+ },
483
+ "device": "cuda",
484
+ "use_amp": false,
485
+ "use_peft": false,
486
+ "push_to_hub": false,
487
+ "repo_id": null,
488
+ "private": null,
489
+ "tags": null,
490
+ "license": null,
491
+ "pretrained_path": null,
492
+ "horizon": 32,
493
+ "n_action_steps": 24,
494
+ "task_conditioning": "embedding",
495
+ "num_tasks": 10,
496
+ "objective": "diffusion",
497
+ "noise_scheduler_type": "DDPM",
498
+ "num_train_timesteps": 100,
499
+ "beta_schedule": "squaredcos_cap_v2",
500
+ "beta_start": 0.0001,
501
+ "beta_end": 0.02,
502
+ "prediction_type": "epsilon",
503
+ "clip_sample": true,
504
+ "clip_sample_range": 1.0,
505
+ "num_inference_steps": null,
506
+ "sigma_min": 0.0,
507
+ "num_integration_steps": 100,
508
+ "integration_method": "euler",
509
+ "timestep_sampling_strategy": "beta",
510
+ "timestep_sampling_s": 0.999,
511
+ "timestep_sampling_alpha": 1.5,
512
+ "timestep_sampling_beta": 1.0,
513
+ "hidden_dim": 512,
514
+ "num_layers": 6,
515
+ "num_heads": 8,
516
+ "dropout": 0.1,
517
+ "use_positional_encoding": false,
518
+ "timestep_embed_dim": 256,
519
+ "use_rope": true,
520
+ "rope_base": 10000.0,
521
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
522
+ "use_separate_rgb_encoder_per_camera": false,
523
+ "vision_encoder_lr_multiplier": 0.1,
524
+ "image_resize_shape": [
525
+ 232,
526
+ 232
527
+ ],
528
+ "image_crop_shape": [
529
+ 224,
530
+ 224
531
+ ],
532
+ "image_crop_is_random": true,
533
+ "text_encoder_name": "openai/clip-vit-base-patch16",
534
+ "tokenizer_max_length": 77,
535
+ "tokenizer_padding": "max_length",
536
+ "tokenizer_padding_side": "right",
537
+ "tokenizer_truncation": true,
538
+ "normalization_mapping": {
539
+ "VISUAL": "MEAN_STD",
540
+ "STATE": "MIN_MAX",
541
+ "ACTION": "MIN_MAX"
542
+ },
543
+ "optimizer_lr": 5e-05,
544
+ "optimizer_betas": [
545
+ 0.95,
546
+ 0.999
547
+ ],
548
+ "optimizer_eps": 1e-08,
549
+ "optimizer_weight_decay": 0.0001,
550
+ "scheduler_name": "cosine",
551
+ "scheduler_warmup_steps": 0,
552
+ "do_mask_loss_for_padding": false,
553
+ "drop_n_last_frames": 7
554
+ },
555
+ "output_dir": "/opt/dlami/nvme/outputs/dit_diffusion_20260511_105440",
556
+ "job_name": "aic_multi_task_dit",
557
+ "resume": false,
558
+ "seed": 42,
559
+ "cudnn_deterministic": false,
560
+ "num_workers": 4,
561
+ "batch_size": 12,
562
+ "steps": 100000,
563
+ "eval_freq": 0,
564
+ "log_freq": 100,
565
+ "tolerance_s": 0.1,
566
+ "save_checkpoint": true,
567
+ "save_freq": 10000,
568
+ "use_policy_training_preset": true,
569
+ "optimizer": {
570
+ "type": "adam",
571
+ "lr": 5e-05,
572
+ "weight_decay": 0.0001,
573
+ "grad_clip_norm": 10.0,
574
+ "betas": [
575
+ 0.95,
576
+ 0.999
577
+ ],
578
+ "eps": 1e-08
579
+ },
580
+ "scheduler": {
581
+ "type": "diffuser",
582
+ "num_warmup_steps": 0,
583
+ "name": "cosine"
584
+ },
585
+ "eval": {
586
+ "n_episodes": 50,
587
+ "batch_size": 50,
588
+ "use_async_envs": false
589
+ },
590
+ "wandb": {
591
+ "enable": true,
592
+ "disable_artifact": false,
593
+ "project": "aic_sfp_dit_diffusion_embedding",
594
+ "entity": "gustave-charles-s-eth-z-rich",
595
+ "notes": null,
596
+ "run_id": "2ipoouae",
597
+ "mode": null,
598
+ "add_tags": true
599
+ },
600
+ "peft": null,
601
+ "use_rabc": false,
602
+ "rabc_progress_path": null,
603
+ "rabc_kappa": 0.01,
604
+ "rabc_epsilon": 1e-06,
605
+ "rabc_head_mode": "sparse",
606
+ "rename_map": {},
607
+ "checkpoint_path": null
608
+ }
checkpoints/030000/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,308 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 3.969463130731183e-05,
4
+ "betas": [
5
+ 0.95,
6
+ 0.999
7
+ ],
8
+ "eps": 1e-08,
9
+ "weight_decay": 0.0001,
10
+ "amsgrad": false,
11
+ "maximize": false,
12
+ "foreach": null,
13
+ "capturable": false,
14
+ "differentiable": false,
15
+ "fused": null,
16
+ "decoupled_weight_decay": false,
17
+ "initial_lr": 5e-05,
18
+ "params": [
19
+ 0,
20
+ 1,
21
+ 2,
22
+ 3,
23
+ 4,
24
+ 5,
25
+ 6,
26
+ 7,
27
+ 8,
28
+ 9,
29
+ 10,
30
+ 11,
31
+ 12,
32
+ 13,
33
+ 14,
34
+ 15,
35
+ 16,
36
+ 17,
37
+ 18,
38
+ 19,
39
+ 20,
40
+ 21,
41
+ 22,
42
+ 23,
43
+ 24,
44
+ 25,
45
+ 26,
46
+ 27,
47
+ 28,
48
+ 29,
49
+ 30,
50
+ 31,
51
+ 32,
52
+ 33,
53
+ 34,
54
+ 35,
55
+ 36,
56
+ 37,
57
+ 38,
58
+ 39,
59
+ 40,
60
+ 41,
61
+ 42,
62
+ 43,
63
+ 44,
64
+ 45,
65
+ 46,
66
+ 47,
67
+ 48,
68
+ 49,
69
+ 50,
70
+ 51,
71
+ 52,
72
+ 53,
73
+ 54,
74
+ 55,
75
+ 56,
76
+ 57,
77
+ 58,
78
+ 59,
79
+ 60,
80
+ 61,
81
+ 62,
82
+ 63,
83
+ 64,
84
+ 65,
85
+ 66,
86
+ 67,
87
+ 68
88
+ ]
89
+ },
90
+ {
91
+ "lr": 3.969463130731183e-06,
92
+ "betas": [
93
+ 0.95,
94
+ 0.999
95
+ ],
96
+ "eps": 1e-08,
97
+ "weight_decay": 0.0001,
98
+ "amsgrad": false,
99
+ "maximize": false,
100
+ "foreach": null,
101
+ "capturable": false,
102
+ "differentiable": false,
103
+ "fused": null,
104
+ "decoupled_weight_decay": false,
105
+ "initial_lr": 5e-06,
106
+ "params": [
107
+ 69,
108
+ 70,
109
+ 71,
110
+ 72,
111
+ 73,
112
+ 74,
113
+ 75,
114
+ 76,
115
+ 77,
116
+ 78,
117
+ 79,
118
+ 80,
119
+ 81,
120
+ 82,
121
+ 83,
122
+ 84,
123
+ 85,
124
+ 86,
125
+ 87,
126
+ 88,
127
+ 89,
128
+ 90,
129
+ 91,
130
+ 92,
131
+ 93,
132
+ 94,
133
+ 95,
134
+ 96,
135
+ 97,
136
+ 98,
137
+ 99,
138
+ 100,
139
+ 101,
140
+ 102,
141
+ 103,
142
+ 104,
143
+ 105,
144
+ 106,
145
+ 107,
146
+ 108,
147
+ 109,
148
+ 110,
149
+ 111,
150
+ 112,
151
+ 113,
152
+ 114,
153
+ 115,
154
+ 116,
155
+ 117,
156
+ 118,
157
+ 119,
158
+ 120,
159
+ 121,
160
+ 122,
161
+ 123,
162
+ 124,
163
+ 125,
164
+ 126,
165
+ 127,
166
+ 128,
167
+ 129,
168
+ 130,
169
+ 131,
170
+ 132,
171
+ 133,
172
+ 134,
173
+ 135,
174
+ 136,
175
+ 137,
176
+ 138,
177
+ 139,
178
+ 140,
179
+ 141,
180
+ 142,
181
+ 143,
182
+ 144,
183
+ 145,
184
+ 146,
185
+ 147,
186
+ 148,
187
+ 149,
188
+ 150,
189
+ 151,
190
+ 152,
191
+ 153,
192
+ 154,
193
+ 155,
194
+ 156,
195
+ 157,
196
+ 158,
197
+ 159,
198
+ 160,
199
+ 161,
200
+ 162,
201
+ 163,
202
+ 164,
203
+ 165,
204
+ 166,
205
+ 167,
206
+ 168,
207
+ 169,
208
+ 170,
209
+ 171,
210
+ 172,
211
+ 173,
212
+ 174,
213
+ 175,
214
+ 176,
215
+ 177,
216
+ 178,
217
+ 179,
218
+ 180,
219
+ 181,
220
+ 182,
221
+ 183,
222
+ 184,
223
+ 185,
224
+ 186,
225
+ 187,
226
+ 188,
227
+ 189,
228
+ 190,
229
+ 191,
230
+ 192,
231
+ 193,
232
+ 194,
233
+ 195,
234
+ 196,
235
+ 197,
236
+ 198,
237
+ 199,
238
+ 200,
239
+ 201,
240
+ 202,
241
+ 203,
242
+ 204,
243
+ 205,
244
+ 206,
245
+ 207,
246
+ 208,
247
+ 209,
248
+ 210,
249
+ 211,
250
+ 212,
251
+ 213,
252
+ 214,
253
+ 215,
254
+ 216,
255
+ 217,
256
+ 218,
257
+ 219,
258
+ 220,
259
+ 221,
260
+ 222,
261
+ 223,
262
+ 224,
263
+ 225,
264
+ 226,
265
+ 227,
266
+ 228,
267
+ 229,
268
+ 230,
269
+ 231,
270
+ 232,
271
+ 233,
272
+ 234,
273
+ 235,
274
+ 236,
275
+ 237,
276
+ 238,
277
+ 239,
278
+ 240,
279
+ 241,
280
+ 242,
281
+ 243,
282
+ 244,
283
+ 245,
284
+ 246,
285
+ 247,
286
+ 248,
287
+ 249,
288
+ 250,
289
+ 251,
290
+ 252,
291
+ 253,
292
+ 254,
293
+ 255,
294
+ 256,
295
+ 257,
296
+ 258,
297
+ 259,
298
+ 260,
299
+ 261,
300
+ 262,
301
+ 263,
302
+ 264,
303
+ 265,
304
+ 266,
305
+ 267
306
+ ]
307
+ }
308
+ ]
checkpoints/030000/training_state/optimizer_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1e5c020708cfa6340f398a6c82644c6deef2a8031586a3173e8b14aa9cff2890
3
+ size 1715911736
checkpoints/030000/training_state/rng_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1feedd4950dab3d6d768e5b3c52f064e92d3f02cac1c128ccae7173324e290b4
3
+ size 15708
checkpoints/030000/training_state/scheduler_state.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_lrs": [
3
+ 5e-05,
4
+ 5e-06
5
+ ],
6
+ "last_epoch": 30000,
7
+ "_step_count": 30001,
8
+ "_get_lr_called_within_step": false,
9
+ "_last_lr": [
10
+ 3.969463130731183e-05,
11
+ 3.969463130731183e-06
12
+ ],
13
+ "lr_lambdas": [
14
+ null,
15
+ null
16
+ ]
17
+ }
checkpoints/030000/training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 30000
3
+ }
checkpoints/040000/pretrained_model/config.json ADDED
@@ -0,0 +1,121 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "aic_multi_task_dit",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.images.left_camera": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 256,
10
+ 288
11
+ ]
12
+ },
13
+ "observation.images.center_camera": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 256,
18
+ 288
19
+ ]
20
+ },
21
+ "observation.images.right_camera": {
22
+ "type": "VISUAL",
23
+ "shape": [
24
+ 3,
25
+ 256,
26
+ 288
27
+ ]
28
+ },
29
+ "observation.state": {
30
+ "type": "STATE",
31
+ "shape": [
32
+ 26
33
+ ]
34
+ },
35
+ "observation.wrench": {
36
+ "type": "STATE",
37
+ "shape": [
38
+ 6
39
+ ]
40
+ }
41
+ },
42
+ "output_features": {
43
+ "action": {
44
+ "type": "ACTION",
45
+ "shape": [
46
+ 7
47
+ ]
48
+ }
49
+ },
50
+ "device": "cuda",
51
+ "use_amp": false,
52
+ "use_peft": false,
53
+ "push_to_hub": false,
54
+ "repo_id": null,
55
+ "private": null,
56
+ "tags": null,
57
+ "license": null,
58
+ "pretrained_path": null,
59
+ "horizon": 32,
60
+ "n_action_steps": 24,
61
+ "task_conditioning": "embedding",
62
+ "num_tasks": 10,
63
+ "objective": "diffusion",
64
+ "noise_scheduler_type": "DDPM",
65
+ "num_train_timesteps": 100,
66
+ "beta_schedule": "squaredcos_cap_v2",
67
+ "beta_start": 0.0001,
68
+ "beta_end": 0.02,
69
+ "prediction_type": "epsilon",
70
+ "clip_sample": true,
71
+ "clip_sample_range": 1.0,
72
+ "num_inference_steps": null,
73
+ "sigma_min": 0.0,
74
+ "num_integration_steps": 100,
75
+ "integration_method": "euler",
76
+ "timestep_sampling_strategy": "beta",
77
+ "timestep_sampling_s": 0.999,
78
+ "timestep_sampling_alpha": 1.5,
79
+ "timestep_sampling_beta": 1.0,
80
+ "hidden_dim": 512,
81
+ "num_layers": 6,
82
+ "num_heads": 8,
83
+ "dropout": 0.1,
84
+ "use_positional_encoding": false,
85
+ "timestep_embed_dim": 256,
86
+ "use_rope": true,
87
+ "rope_base": 10000.0,
88
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
89
+ "use_separate_rgb_encoder_per_camera": false,
90
+ "vision_encoder_lr_multiplier": 0.1,
91
+ "image_resize_shape": [
92
+ 232,
93
+ 232
94
+ ],
95
+ "image_crop_shape": [
96
+ 224,
97
+ 224
98
+ ],
99
+ "image_crop_is_random": true,
100
+ "text_encoder_name": "openai/clip-vit-base-patch16",
101
+ "tokenizer_max_length": 77,
102
+ "tokenizer_padding": "max_length",
103
+ "tokenizer_padding_side": "right",
104
+ "tokenizer_truncation": true,
105
+ "normalization_mapping": {
106
+ "VISUAL": "MEAN_STD",
107
+ "STATE": "MIN_MAX",
108
+ "ACTION": "MIN_MAX"
109
+ },
110
+ "optimizer_lr": 5e-05,
111
+ "optimizer_betas": [
112
+ 0.95,
113
+ 0.999
114
+ ],
115
+ "optimizer_eps": 1e-08,
116
+ "optimizer_weight_decay": 0.0001,
117
+ "scheduler_name": "cosine",
118
+ "scheduler_warmup_steps": 0,
119
+ "do_mask_loss_for_padding": false,
120
+ "drop_n_last_frames": 7
121
+ }
checkpoints/040000/pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4f664e4fd40459da01925cf223624a103ca95a5ae336f3804899bf22aac4b17f
3
+ size 857967220
checkpoints/040000/pretrained_model/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": "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
+ }
checkpoints/040000/pretrained_model/policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d58fa83446e4b0bdfc09a02565d49fca3dd894c73486df2d1ce3a22e752cd6
3
+ size 11368
checkpoints/040000/pretrained_model/policy_preprocessor.json ADDED
@@ -0,0 +1,78 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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.left_camera": {
27
+ "type": "VISUAL",
28
+ "shape": [
29
+ 3,
30
+ 256,
31
+ 288
32
+ ]
33
+ },
34
+ "observation.images.center_camera": {
35
+ "type": "VISUAL",
36
+ "shape": [
37
+ 3,
38
+ 256,
39
+ 288
40
+ ]
41
+ },
42
+ "observation.images.right_camera": {
43
+ "type": "VISUAL",
44
+ "shape": [
45
+ 3,
46
+ 256,
47
+ 288
48
+ ]
49
+ },
50
+ "observation.state": {
51
+ "type": "STATE",
52
+ "shape": [
53
+ 26
54
+ ]
55
+ },
56
+ "observation.wrench": {
57
+ "type": "STATE",
58
+ "shape": [
59
+ 6
60
+ ]
61
+ },
62
+ "action": {
63
+ "type": "ACTION",
64
+ "shape": [
65
+ 7
66
+ ]
67
+ }
68
+ },
69
+ "norm_map": {
70
+ "VISUAL": "MEAN_STD",
71
+ "STATE": "MIN_MAX",
72
+ "ACTION": "MIN_MAX"
73
+ }
74
+ },
75
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
76
+ }
77
+ ]
78
+ }
checkpoints/040000/pretrained_model/policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a9d58fa83446e4b0bdfc09a02565d49fca3dd894c73486df2d1ce3a22e752cd6
3
+ size 11368
checkpoints/040000/pretrained_model/train_config.json ADDED
@@ -0,0 +1,608 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "local/local_aic_cheatcode_demos_dit_snap",
4
+ "root": "/opt/dlami/nvme/lerobot/local_aic_cheatcode_demos_dit_snap",
5
+ "episodes": [
6
+ 3,
7
+ 5,
8
+ 8,
9
+ 157,
10
+ 183,
11
+ 184,
12
+ 185,
13
+ 186,
14
+ 187,
15
+ 188,
16
+ 189,
17
+ 190,
18
+ 192,
19
+ 193,
20
+ 194,
21
+ 195,
22
+ 196,
23
+ 198,
24
+ 199,
25
+ 200,
26
+ 201,
27
+ 203,
28
+ 204,
29
+ 205,
30
+ 206,
31
+ 207,
32
+ 208,
33
+ 209,
34
+ 210,
35
+ 211,
36
+ 213,
37
+ 214,
38
+ 215,
39
+ 216,
40
+ 217,
41
+ 218,
42
+ 219,
43
+ 220,
44
+ 221,
45
+ 222,
46
+ 223,
47
+ 224,
48
+ 225,
49
+ 227,
50
+ 228,
51
+ 229,
52
+ 230,
53
+ 232,
54
+ 233,
55
+ 234,
56
+ 236,
57
+ 238,
58
+ 239,
59
+ 241,
60
+ 242,
61
+ 243,
62
+ 244,
63
+ 245,
64
+ 246,
65
+ 247,
66
+ 248,
67
+ 249,
68
+ 250,
69
+ 251,
70
+ 252,
71
+ 253,
72
+ 255,
73
+ 256,
74
+ 257,
75
+ 258,
76
+ 259,
77
+ 260,
78
+ 262,
79
+ 263,
80
+ 264,
81
+ 265,
82
+ 266,
83
+ 267,
84
+ 268,
85
+ 269,
86
+ 270,
87
+ 271,
88
+ 272,
89
+ 273,
90
+ 274,
91
+ 276,
92
+ 277,
93
+ 278,
94
+ 279,
95
+ 280,
96
+ 281,
97
+ 283,
98
+ 284,
99
+ 285,
100
+ 286,
101
+ 287,
102
+ 288,
103
+ 289,
104
+ 290,
105
+ 291,
106
+ 292,
107
+ 293,
108
+ 294,
109
+ 295,
110
+ 298,
111
+ 299,
112
+ 300,
113
+ 301,
114
+ 302,
115
+ 303,
116
+ 304,
117
+ 305,
118
+ 306,
119
+ 307,
120
+ 308,
121
+ 309,
122
+ 310,
123
+ 311,
124
+ 312,
125
+ 313,
126
+ 314,
127
+ 315,
128
+ 316,
129
+ 317,
130
+ 318,
131
+ 319,
132
+ 321,
133
+ 322,
134
+ 323,
135
+ 325,
136
+ 327,
137
+ 328,
138
+ 329,
139
+ 330,
140
+ 332,
141
+ 333,
142
+ 334,
143
+ 335,
144
+ 336,
145
+ 337,
146
+ 338,
147
+ 340,
148
+ 341,
149
+ 342,
150
+ 343,
151
+ 344,
152
+ 345,
153
+ 346,
154
+ 347,
155
+ 348,
156
+ 350,
157
+ 351,
158
+ 352,
159
+ 354,
160
+ 355,
161
+ 356,
162
+ 358,
163
+ 359,
164
+ 360,
165
+ 361,
166
+ 362,
167
+ 363,
168
+ 364,
169
+ 366,
170
+ 367,
171
+ 368,
172
+ 369,
173
+ 370,
174
+ 371,
175
+ 372,
176
+ 373,
177
+ 374,
178
+ 375,
179
+ 377,
180
+ 378,
181
+ 379,
182
+ 380,
183
+ 381,
184
+ 384,
185
+ 385,
186
+ 386,
187
+ 387,
188
+ 388,
189
+ 389,
190
+ 390,
191
+ 391,
192
+ 393,
193
+ 395,
194
+ 397,
195
+ 399,
196
+ 400,
197
+ 401,
198
+ 402,
199
+ 404,
200
+ 405,
201
+ 407,
202
+ 408,
203
+ 409,
204
+ 410,
205
+ 411,
206
+ 413,
207
+ 414,
208
+ 415,
209
+ 416,
210
+ 417,
211
+ 419,
212
+ 420,
213
+ 421,
214
+ 422,
215
+ 423,
216
+ 425,
217
+ 430,
218
+ 431,
219
+ 433,
220
+ 434,
221
+ 436,
222
+ 437,
223
+ 439,
224
+ 440,
225
+ 442,
226
+ 444,
227
+ 447,
228
+ 449,
229
+ 450,
230
+ 452,
231
+ 453,
232
+ 454,
233
+ 455,
234
+ 457,
235
+ 458,
236
+ 460,
237
+ 462,
238
+ 463,
239
+ 464,
240
+ 465,
241
+ 466,
242
+ 467,
243
+ 470,
244
+ 471,
245
+ 473,
246
+ 474,
247
+ 475,
248
+ 476,
249
+ 478,
250
+ 479,
251
+ 480,
252
+ 482,
253
+ 485,
254
+ 490,
255
+ 491,
256
+ 492,
257
+ 497,
258
+ 498,
259
+ 500,
260
+ 501,
261
+ 502,
262
+ 504,
263
+ 507,
264
+ 508,
265
+ 509,
266
+ 512,
267
+ 514,
268
+ 515,
269
+ 516,
270
+ 517,
271
+ 518,
272
+ 520,
273
+ 522,
274
+ 524,
275
+ 527,
276
+ 529,
277
+ 530,
278
+ 532,
279
+ 533,
280
+ 534,
281
+ 535,
282
+ 537,
283
+ 538,
284
+ 539,
285
+ 540,
286
+ 541,
287
+ 542,
288
+ 543,
289
+ 544,
290
+ 546,
291
+ 548,
292
+ 549,
293
+ 550,
294
+ 551,
295
+ 553,
296
+ 555,
297
+ 556,
298
+ 557,
299
+ 558,
300
+ 559,
301
+ 560,
302
+ 562,
303
+ 563,
304
+ 565,
305
+ 566,
306
+ 569,
307
+ 570,
308
+ 571,
309
+ 572,
310
+ 575,
311
+ 576,
312
+ 577,
313
+ 578,
314
+ 580,
315
+ 581,
316
+ 582,
317
+ 584,
318
+ 585,
319
+ 586,
320
+ 587,
321
+ 588,
322
+ 589,
323
+ 590,
324
+ 591,
325
+ 593,
326
+ 594,
327
+ 595,
328
+ 596,
329
+ 598,
330
+ 600,
331
+ 601,
332
+ 602,
333
+ 603,
334
+ 606,
335
+ 607,
336
+ 608,
337
+ 609,
338
+ 612,
339
+ 613,
340
+ 615,
341
+ 616,
342
+ 617,
343
+ 619,
344
+ 620,
345
+ 621,
346
+ 623,
347
+ 624,
348
+ 627,
349
+ 628,
350
+ 631,
351
+ 632,
352
+ 633,
353
+ 634,
354
+ 635,
355
+ 636
356
+ ],
357
+ "image_transforms": {
358
+ "enable": true,
359
+ "max_num_transforms": 3,
360
+ "random_order": true,
361
+ "tfs": {
362
+ "brightness": {
363
+ "weight": 1.0,
364
+ "type": "ColorJitter",
365
+ "kwargs": {
366
+ "brightness": [
367
+ 0.9,
368
+ 1.1
369
+ ]
370
+ }
371
+ },
372
+ "contrast": {
373
+ "weight": 1.0,
374
+ "type": "ColorJitter",
375
+ "kwargs": {
376
+ "contrast": [
377
+ 0.9,
378
+ 1.1
379
+ ]
380
+ }
381
+ },
382
+ "saturation": {
383
+ "weight": 1.0,
384
+ "type": "ColorJitter",
385
+ "kwargs": {
386
+ "saturation": [
387
+ 0.95,
388
+ 1.05
389
+ ]
390
+ }
391
+ },
392
+ "hue": {
393
+ "weight": 1.0,
394
+ "type": "ColorJitter",
395
+ "kwargs": {
396
+ "hue": [
397
+ -0.02,
398
+ 0.02
399
+ ]
400
+ }
401
+ },
402
+ "sharpness": {
403
+ "weight": 1.0,
404
+ "type": "SharpnessJitter",
405
+ "kwargs": {
406
+ "sharpness": [
407
+ 0.9,
408
+ 1.1
409
+ ]
410
+ }
411
+ },
412
+ "affine": {
413
+ "weight": 0.0,
414
+ "type": "RandomAffine",
415
+ "kwargs": {
416
+ "degrees": [
417
+ -5.0,
418
+ 5.0
419
+ ],
420
+ "translate": [
421
+ 0.05,
422
+ 0.05
423
+ ]
424
+ }
425
+ }
426
+ }
427
+ },
428
+ "revision": null,
429
+ "use_imagenet_stats": true,
430
+ "video_backend": "pyav",
431
+ "streaming": false
432
+ },
433
+ "env": null,
434
+ "policy": {
435
+ "type": "aic_multi_task_dit",
436
+ "n_obs_steps": 2,
437
+ "input_features": {
438
+ "observation.images.left_camera": {
439
+ "type": "VISUAL",
440
+ "shape": [
441
+ 3,
442
+ 256,
443
+ 288
444
+ ]
445
+ },
446
+ "observation.images.center_camera": {
447
+ "type": "VISUAL",
448
+ "shape": [
449
+ 3,
450
+ 256,
451
+ 288
452
+ ]
453
+ },
454
+ "observation.images.right_camera": {
455
+ "type": "VISUAL",
456
+ "shape": [
457
+ 3,
458
+ 256,
459
+ 288
460
+ ]
461
+ },
462
+ "observation.state": {
463
+ "type": "STATE",
464
+ "shape": [
465
+ 26
466
+ ]
467
+ },
468
+ "observation.wrench": {
469
+ "type": "STATE",
470
+ "shape": [
471
+ 6
472
+ ]
473
+ }
474
+ },
475
+ "output_features": {
476
+ "action": {
477
+ "type": "ACTION",
478
+ "shape": [
479
+ 7
480
+ ]
481
+ }
482
+ },
483
+ "device": "cuda",
484
+ "use_amp": false,
485
+ "use_peft": false,
486
+ "push_to_hub": false,
487
+ "repo_id": null,
488
+ "private": null,
489
+ "tags": null,
490
+ "license": null,
491
+ "pretrained_path": null,
492
+ "horizon": 32,
493
+ "n_action_steps": 24,
494
+ "task_conditioning": "embedding",
495
+ "num_tasks": 10,
496
+ "objective": "diffusion",
497
+ "noise_scheduler_type": "DDPM",
498
+ "num_train_timesteps": 100,
499
+ "beta_schedule": "squaredcos_cap_v2",
500
+ "beta_start": 0.0001,
501
+ "beta_end": 0.02,
502
+ "prediction_type": "epsilon",
503
+ "clip_sample": true,
504
+ "clip_sample_range": 1.0,
505
+ "num_inference_steps": null,
506
+ "sigma_min": 0.0,
507
+ "num_integration_steps": 100,
508
+ "integration_method": "euler",
509
+ "timestep_sampling_strategy": "beta",
510
+ "timestep_sampling_s": 0.999,
511
+ "timestep_sampling_alpha": 1.5,
512
+ "timestep_sampling_beta": 1.0,
513
+ "hidden_dim": 512,
514
+ "num_layers": 6,
515
+ "num_heads": 8,
516
+ "dropout": 0.1,
517
+ "use_positional_encoding": false,
518
+ "timestep_embed_dim": 256,
519
+ "use_rope": true,
520
+ "rope_base": 10000.0,
521
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
522
+ "use_separate_rgb_encoder_per_camera": false,
523
+ "vision_encoder_lr_multiplier": 0.1,
524
+ "image_resize_shape": [
525
+ 232,
526
+ 232
527
+ ],
528
+ "image_crop_shape": [
529
+ 224,
530
+ 224
531
+ ],
532
+ "image_crop_is_random": true,
533
+ "text_encoder_name": "openai/clip-vit-base-patch16",
534
+ "tokenizer_max_length": 77,
535
+ "tokenizer_padding": "max_length",
536
+ "tokenizer_padding_side": "right",
537
+ "tokenizer_truncation": true,
538
+ "normalization_mapping": {
539
+ "VISUAL": "MEAN_STD",
540
+ "STATE": "MIN_MAX",
541
+ "ACTION": "MIN_MAX"
542
+ },
543
+ "optimizer_lr": 5e-05,
544
+ "optimizer_betas": [
545
+ 0.95,
546
+ 0.999
547
+ ],
548
+ "optimizer_eps": 1e-08,
549
+ "optimizer_weight_decay": 0.0001,
550
+ "scheduler_name": "cosine",
551
+ "scheduler_warmup_steps": 0,
552
+ "do_mask_loss_for_padding": false,
553
+ "drop_n_last_frames": 7
554
+ },
555
+ "output_dir": "/opt/dlami/nvme/outputs/dit_diffusion_20260511_105440",
556
+ "job_name": "aic_multi_task_dit",
557
+ "resume": false,
558
+ "seed": 42,
559
+ "cudnn_deterministic": false,
560
+ "num_workers": 4,
561
+ "batch_size": 12,
562
+ "steps": 100000,
563
+ "eval_freq": 0,
564
+ "log_freq": 100,
565
+ "tolerance_s": 0.1,
566
+ "save_checkpoint": true,
567
+ "save_freq": 10000,
568
+ "use_policy_training_preset": true,
569
+ "optimizer": {
570
+ "type": "adam",
571
+ "lr": 5e-05,
572
+ "weight_decay": 0.0001,
573
+ "grad_clip_norm": 10.0,
574
+ "betas": [
575
+ 0.95,
576
+ 0.999
577
+ ],
578
+ "eps": 1e-08
579
+ },
580
+ "scheduler": {
581
+ "type": "diffuser",
582
+ "num_warmup_steps": 0,
583
+ "name": "cosine"
584
+ },
585
+ "eval": {
586
+ "n_episodes": 50,
587
+ "batch_size": 50,
588
+ "use_async_envs": false
589
+ },
590
+ "wandb": {
591
+ "enable": true,
592
+ "disable_artifact": false,
593
+ "project": "aic_sfp_dit_diffusion_embedding",
594
+ "entity": "gustave-charles-s-eth-z-rich",
595
+ "notes": null,
596
+ "run_id": "2ipoouae",
597
+ "mode": null,
598
+ "add_tags": true
599
+ },
600
+ "peft": null,
601
+ "use_rabc": false,
602
+ "rabc_progress_path": null,
603
+ "rabc_kappa": 0.01,
604
+ "rabc_epsilon": 1e-06,
605
+ "rabc_head_mode": "sparse",
606
+ "rename_map": {},
607
+ "checkpoint_path": null
608
+ }
checkpoints/040000/training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,308 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "lr": 3.272542485937369e-05,
4
+ "betas": [
5
+ 0.95,
6
+ 0.999
7
+ ],
8
+ "eps": 1e-08,
9
+ "weight_decay": 0.0001,
10
+ "amsgrad": false,
11
+ "maximize": false,
12
+ "foreach": null,
13
+ "capturable": false,
14
+ "differentiable": false,
15
+ "fused": null,
16
+ "decoupled_weight_decay": false,
17
+ "initial_lr": 5e-05,
18
+ "params": [
19
+ 0,
20
+ 1,
21
+ 2,
22
+ 3,
23
+ 4,
24
+ 5,
25
+ 6,
26
+ 7,
27
+ 8,
28
+ 9,
29
+ 10,
30
+ 11,
31
+ 12,
32
+ 13,
33
+ 14,
34
+ 15,
35
+ 16,
36
+ 17,
37
+ 18,
38
+ 19,
39
+ 20,
40
+ 21,
41
+ 22,
42
+ 23,
43
+ 24,
44
+ 25,
45
+ 26,
46
+ 27,
47
+ 28,
48
+ 29,
49
+ 30,
50
+ 31,
51
+ 32,
52
+ 33,
53
+ 34,
54
+ 35,
55
+ 36,
56
+ 37,
57
+ 38,
58
+ 39,
59
+ 40,
60
+ 41,
61
+ 42,
62
+ 43,
63
+ 44,
64
+ 45,
65
+ 46,
66
+ 47,
67
+ 48,
68
+ 49,
69
+ 50,
70
+ 51,
71
+ 52,
72
+ 53,
73
+ 54,
74
+ 55,
75
+ 56,
76
+ 57,
77
+ 58,
78
+ 59,
79
+ 60,
80
+ 61,
81
+ 62,
82
+ 63,
83
+ 64,
84
+ 65,
85
+ 66,
86
+ 67,
87
+ 68
88
+ ]
89
+ },
90
+ {
91
+ "lr": 3.272542485937369e-06,
92
+ "betas": [
93
+ 0.95,
94
+ 0.999
95
+ ],
96
+ "eps": 1e-08,
97
+ "weight_decay": 0.0001,
98
+ "amsgrad": false,
99
+ "maximize": false,
100
+ "foreach": null,
101
+ "capturable": false,
102
+ "differentiable": false,
103
+ "fused": null,
104
+ "decoupled_weight_decay": false,
105
+ "initial_lr": 5e-06,
106
+ "params": [
107
+ 69,
108
+ 70,
109
+ 71,
110
+ 72,
111
+ 73,
112
+ 74,
113
+ 75,
114
+ 76,
115
+ 77,
116
+ 78,
117
+ 79,
118
+ 80,
119
+ 81,
120
+ 82,
121
+ 83,
122
+ 84,
123
+ 85,
124
+ 86,
125
+ 87,
126
+ 88,
127
+ 89,
128
+ 90,
129
+ 91,
130
+ 92,
131
+ 93,
132
+ 94,
133
+ 95,
134
+ 96,
135
+ 97,
136
+ 98,
137
+ 99,
138
+ 100,
139
+ 101,
140
+ 102,
141
+ 103,
142
+ 104,
143
+ 105,
144
+ 106,
145
+ 107,
146
+ 108,
147
+ 109,
148
+ 110,
149
+ 111,
150
+ 112,
151
+ 113,
152
+ 114,
153
+ 115,
154
+ 116,
155
+ 117,
156
+ 118,
157
+ 119,
158
+ 120,
159
+ 121,
160
+ 122,
161
+ 123,
162
+ 124,
163
+ 125,
164
+ 126,
165
+ 127,
166
+ 128,
167
+ 129,
168
+ 130,
169
+ 131,
170
+ 132,
171
+ 133,
172
+ 134,
173
+ 135,
174
+ 136,
175
+ 137,
176
+ 138,
177
+ 139,
178
+ 140,
179
+ 141,
180
+ 142,
181
+ 143,
182
+ 144,
183
+ 145,
184
+ 146,
185
+ 147,
186
+ 148,
187
+ 149,
188
+ 150,
189
+ 151,
190
+ 152,
191
+ 153,
192
+ 154,
193
+ 155,
194
+ 156,
195
+ 157,
196
+ 158,
197
+ 159,
198
+ 160,
199
+ 161,
200
+ 162,
201
+ 163,
202
+ 164,
203
+ 165,
204
+ 166,
205
+ 167,
206
+ 168,
207
+ 169,
208
+ 170,
209
+ 171,
210
+ 172,
211
+ 173,
212
+ 174,
213
+ 175,
214
+ 176,
215
+ 177,
216
+ 178,
217
+ 179,
218
+ 180,
219
+ 181,
220
+ 182,
221
+ 183,
222
+ 184,
223
+ 185,
224
+ 186,
225
+ 187,
226
+ 188,
227
+ 189,
228
+ 190,
229
+ 191,
230
+ 192,
231
+ 193,
232
+ 194,
233
+ 195,
234
+ 196,
235
+ 197,
236
+ 198,
237
+ 199,
238
+ 200,
239
+ 201,
240
+ 202,
241
+ 203,
242
+ 204,
243
+ 205,
244
+ 206,
245
+ 207,
246
+ 208,
247
+ 209,
248
+ 210,
249
+ 211,
250
+ 212,
251
+ 213,
252
+ 214,
253
+ 215,
254
+ 216,
255
+ 217,
256
+ 218,
257
+ 219,
258
+ 220,
259
+ 221,
260
+ 222,
261
+ 223,
262
+ 224,
263
+ 225,
264
+ 226,
265
+ 227,
266
+ 228,
267
+ 229,
268
+ 230,
269
+ 231,
270
+ 232,
271
+ 233,
272
+ 234,
273
+ 235,
274
+ 236,
275
+ 237,
276
+ 238,
277
+ 239,
278
+ 240,
279
+ 241,
280
+ 242,
281
+ 243,
282
+ 244,
283
+ 245,
284
+ 246,
285
+ 247,
286
+ 248,
287
+ 249,
288
+ 250,
289
+ 251,
290
+ 252,
291
+ 253,
292
+ 254,
293
+ 255,
294
+ 256,
295
+ 257,
296
+ 258,
297
+ 259,
298
+ 260,
299
+ 261,
300
+ 262,
301
+ 263,
302
+ 264,
303
+ 265,
304
+ 266,
305
+ 267
306
+ ]
307
+ }
308
+ ]
checkpoints/040000/training_state/optimizer_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2ed5bca0e31ac5476ac4778fac17f1bee70f4748b4323eb888309d26959b924
3
+ size 1715911736
checkpoints/040000/training_state/rng_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a07ba24ed79df5ed4d2d78e7b5d3aaac90d64676ce9d82e51931397435fcdd10
3
+ size 15708
checkpoints/040000/training_state/scheduler_state.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_lrs": [
3
+ 5e-05,
4
+ 5e-06
5
+ ],
6
+ "last_epoch": 40000,
7
+ "_step_count": 40001,
8
+ "_get_lr_called_within_step": false,
9
+ "_last_lr": [
10
+ 3.272542485937369e-05,
11
+ 3.272542485937369e-06
12
+ ],
13
+ "lr_lambdas": [
14
+ null,
15
+ null
16
+ ]
17
+ }
checkpoints/040000/training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 40000
3
+ }
checkpoints/050000/pretrained_model/config.json ADDED
@@ -0,0 +1,121 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "aic_multi_task_dit",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.images.left_camera": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 256,
10
+ 288
11
+ ]
12
+ },
13
+ "observation.images.center_camera": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 256,
18
+ 288
19
+ ]
20
+ },
21
+ "observation.images.right_camera": {
22
+ "type": "VISUAL",
23
+ "shape": [
24
+ 3,
25
+ 256,
26
+ 288
27
+ ]
28
+ },
29
+ "observation.state": {
30
+ "type": "STATE",
31
+ "shape": [
32
+ 26
33
+ ]
34
+ },
35
+ "observation.wrench": {
36
+ "type": "STATE",
37
+ "shape": [
38
+ 6
39
+ ]
40
+ }
41
+ },
42
+ "output_features": {
43
+ "action": {
44
+ "type": "ACTION",
45
+ "shape": [
46
+ 7
47
+ ]
48
+ }
49
+ },
50
+ "device": "cuda",
51
+ "use_amp": false,
52
+ "use_peft": false,
53
+ "push_to_hub": false,
54
+ "repo_id": null,
55
+ "private": null,
56
+ "tags": null,
57
+ "license": null,
58
+ "pretrained_path": null,
59
+ "horizon": 32,
60
+ "n_action_steps": 24,
61
+ "task_conditioning": "embedding",
62
+ "num_tasks": 10,
63
+ "objective": "diffusion",
64
+ "noise_scheduler_type": "DDPM",
65
+ "num_train_timesteps": 100,
66
+ "beta_schedule": "squaredcos_cap_v2",
67
+ "beta_start": 0.0001,
68
+ "beta_end": 0.02,
69
+ "prediction_type": "epsilon",
70
+ "clip_sample": true,
71
+ "clip_sample_range": 1.0,
72
+ "num_inference_steps": null,
73
+ "sigma_min": 0.0,
74
+ "num_integration_steps": 100,
75
+ "integration_method": "euler",
76
+ "timestep_sampling_strategy": "beta",
77
+ "timestep_sampling_s": 0.999,
78
+ "timestep_sampling_alpha": 1.5,
79
+ "timestep_sampling_beta": 1.0,
80
+ "hidden_dim": 512,
81
+ "num_layers": 6,
82
+ "num_heads": 8,
83
+ "dropout": 0.1,
84
+ "use_positional_encoding": false,
85
+ "timestep_embed_dim": 256,
86
+ "use_rope": true,
87
+ "rope_base": 10000.0,
88
+ "vision_encoder_name": "openai/clip-vit-base-patch16",
89
+ "use_separate_rgb_encoder_per_camera": false,
90
+ "vision_encoder_lr_multiplier": 0.1,
91
+ "image_resize_shape": [
92
+ 232,
93
+ 232
94
+ ],
95
+ "image_crop_shape": [
96
+ 224,
97
+ 224
98
+ ],
99
+ "image_crop_is_random": true,
100
+ "text_encoder_name": "openai/clip-vit-base-patch16",
101
+ "tokenizer_max_length": 77,
102
+ "tokenizer_padding": "max_length",
103
+ "tokenizer_padding_side": "right",
104
+ "tokenizer_truncation": true,
105
+ "normalization_mapping": {
106
+ "VISUAL": "MEAN_STD",
107
+ "STATE": "MIN_MAX",
108
+ "ACTION": "MIN_MAX"
109
+ },
110
+ "optimizer_lr": 5e-05,
111
+ "optimizer_betas": [
112
+ 0.95,
113
+ 0.999
114
+ ],
115
+ "optimizer_eps": 1e-08,
116
+ "optimizer_weight_decay": 0.0001,
117
+ "scheduler_name": "cosine",
118
+ "scheduler_warmup_steps": 0,
119
+ "do_mask_loss_for_padding": false,
120
+ "drop_n_last_frames": 7
121
+ }
checkpoints/050000/pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2dd574d9991fe7ef967f339d64f53e5072ebb9143625d43a9bad0c298aeae14
3
+ size 857967220