Hugging Face's logo

m1ku2
/
pack3_base

Model card Files
xet
 Community
pack3_base
File size: 8,071 Bytes
1479675
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
output_dir: ./runs/pack_3_objects_plus_base/2026-06-08_23-08-07
batch_size: 8
num_workers: 2
lr_scheduler_type: cosine
learning_rate: 0.0001
num_epochs: 100
max_steps: 20000
log_every: 10
save_every: 2000
eval_every: 500
eval_num_inference_steps: 10
gradient_accumulation_steps: 1
mixed_precision: bf16
seed: 42
max_grad_norm: 1.0
weight_decay: 0.01
resume: null
partial_resume: null
wandb:
  enabled: true
  workspace: null
  project: fast-wam-real
  name: pack_3_objects_plus_base
  group: null
  mode: online
data:
  train:
    _target_: fastwam.datasets.lerobot.robot_video_dataset.RobotVideoDataset
    dataset_dirs:
    - ./data/pack_3_objects_plus/perfect_lerobot
    shape_meta:
      images:
      - key: cam_high
        raw_shape:
        - 3
        - 480
        - 640
        shape:
        - 3
        - 240
        - 320
      - key: cam_left_wrist
        raw_shape:
        - 3
        - 480
        - 640
        shape:
        - 3
        - 240
        - 320
      - key: cam_right_wrist
        raw_shape:
        - 3
        - 480
        - 640
        shape:
        - 3
        - 240
        - 320
      action:
      - key: default
        raw_shape: 14
        shape: 14
      state:
      - key: default
        raw_shape: 14
        shape: 14
    num_frames: 33
    global_sample_stride: 1
    action_video_freq_ratio: 4
    video_size:
    - 384
    - 320
    camera_key: null
    val_set_proportion: 0.01
    is_training_set: true
    pretrained_norm_stats: null
    skip_padding_as_possible: false
    concat_multi_camera: robotwin
    processor:
      _target_: fastwam.datasets.lerobot.processors.fastwam_processor.FastWAMProcessor
      shape_meta:
        images:
        - key: cam_high
          raw_shape:
          - 3
          - 480
          - 640
          shape:
          - 3
          - 240
          - 320
        - key: cam_left_wrist
          raw_shape:
          - 3
          - 480
          - 640
          shape:
          - 3
          - 240
          - 320
        - key: cam_right_wrist
          raw_shape:
          - 3
          - 480
          - 640
          shape:
          - 3
          - 240
          - 320
        action:
        - key: default
          raw_shape: 14
          shape: 14
        state:
        - key: default
          raw_shape: 14
          shape: 14
      num_obs_steps: 33
      num_output_cameras: 3
      action_output_dim: 14
      proprio_output_dim: 14
      action_state_transforms: null
      use_stepwise_action_norm: false
      norm_default_mode: z-score
      norm_exception_mode: null
      action_state_merger:
        _target_: fastwam.datasets.lerobot.transforms.action_state_merger.ConcatLeftAlign
      train_transforms:
      - _target_: fastwam.datasets.lerobot.transforms.image.ToTensor
      - _target_: torchvision.transforms.Resize
        size:
        - 240
        - 320
      val_transforms:
      - _target_: fastwam.datasets.lerobot.transforms.image.ToTensor
      - _target_: torchvision.transforms.Resize
        size:
        - 240
        - 320
      image_augmentation: true
      train_aug_transforms:
        cam_high:
        - _target_: torchvision.transforms.RandomCrop
          size:
          - 228
          - 304
        - _target_: torchvision.transforms.Resize
          size:
          - 240
          - 320
        - _target_: torchvision.transforms.RandomRotation
          degrees: 5
        - _target_: torchvision.transforms.ColorJitter
          brightness: 0.3
          contrast: 0.4
          saturation: 0.5
        cam_left_wrist:
        - _target_: torchvision.transforms.ColorJitter
          brightness: 0.3
          contrast: 0.4
          saturation: 0.5
        cam_right_wrist:
        - _target_: torchvision.transforms.ColorJitter
          brightness: 0.3
          contrast: 0.4
          saturation: 0.5
    text_embedding_cache_dir: ./data/text_embeds_cache/pack_3_objects_plus
    context_len: 128
  val:
    _target_: fastwam.datasets.lerobot.robot_video_dataset.RobotVideoDataset
    dataset_dirs:
    - ./data/pack_3_objects_plus/perfect_lerobot
    shape_meta:
      images:
      - key: cam_high
        raw_shape:
        - 3
        - 480
        - 640
        shape:
        - 3
        - 240
        - 320
      - key: cam_left_wrist
        raw_shape:
        - 3
        - 480
        - 640
        shape:
        - 3
        - 240
        - 320
      - key: cam_right_wrist
        raw_shape:
        - 3
        - 480
        - 640
        shape:
        - 3
        - 240
        - 320
      action:
      - key: default
        raw_shape: 14
        shape: 14
      state:
      - key: default
        raw_shape: 14
        shape: 14
    num_frames: 33
    global_sample_stride: 1
    action_video_freq_ratio: 4
    video_size:
    - 384
    - 320
    camera_key: null
    val_set_proportion: 0.01
    is_training_set: false
    pretrained_norm_stats: null
    skip_padding_as_possible: false
    concat_multi_camera: robotwin
    processor:
      _target_: fastwam.datasets.lerobot.processors.fastwam_processor.FastWAMProcessor
      shape_meta:
        images:
        - key: cam_high
          raw_shape:
          - 3
          - 480
          - 640
          shape:
          - 3
          - 240
          - 320
        - key: cam_left_wrist
          raw_shape:
          - 3
          - 480
          - 640
          shape:
          - 3
          - 240
          - 320
        - key: cam_right_wrist
          raw_shape:
          - 3
          - 480
          - 640
          shape:
          - 3
          - 240
          - 320
        action:
        - key: default
          raw_shape: 14
          shape: 14
        state:
        - key: default
          raw_shape: 14
          shape: 14
      num_obs_steps: 33
      num_output_cameras: 3
      action_output_dim: 14
      proprio_output_dim: 14
      action_state_transforms: null
      use_stepwise_action_norm: false
      norm_default_mode: z-score
      norm_exception_mode: null
      action_state_merger:
        _target_: fastwam.datasets.lerobot.transforms.action_state_merger.ConcatLeftAlign
      train_transforms:
      - _target_: fastwam.datasets.lerobot.transforms.image.ToTensor
      - _target_: torchvision.transforms.Resize
        size:
        - 240
        - 320
      val_transforms:
      - _target_: fastwam.datasets.lerobot.transforms.image.ToTensor
      - _target_: torchvision.transforms.Resize
        size:
        - 240
        - 320
    text_embedding_cache_dir: ./data/text_embeds_cache/pack_3_objects_plus
    context_len: 128
model:
  _target_: fastwam.runtime.create_fastwam
  model_id: Wan-AI/Wan2.2-TI2V-5B
  tokenizer_model_id: Wan-AI/Wan2.2-TI2V-5B
  dit_checkpoint_path: null
  tokenizer_max_len: 128
  load_text_encoder: false
  proprio_dim: 14
  redirect_common_files: false
  mot_checkpoint_mixed_attn: false
  action_dit_pretrained_path: checkpoints/ActionDiT_linear_interp_Wan22_alphascale_1024hdim.pt
  skip_dit_load_from_pretrain: false
  video_dit_config:
    has_image_input: false
    patch_size:
    - 1
    - 2
    - 2
    in_dim: 48
    hidden_dim: 3072
    ffn_dim: 14336
    freq_dim: 256
    text_dim: 4096
    out_dim: 48
    num_heads: 24
    attn_head_dim: 128
    num_layers: 30
    eps: 1.0e-06
    seperated_timestep: true
    require_clip_embedding: false
    require_vae_embedding: false
    fuse_vae_embedding_in_latents: true
    use_gradient_checkpointing: false
    video_attention_mask_mode: first_frame_causal
    action_conditioned: false
    action_dim: 14
    action_group_causal_mask_mode: group_diagonal
  action_dit_config:
    action_dim: 14
    hidden_dim: 1024
    ffn_dim: 4096
    num_heads: 24
    attn_head_dim: 128
    num_layers: 30
    text_dim: 4096
    freq_dim: 256
    eps: 1.0e-06
    use_gradient_checkpointing: false
  video_scheduler:
    train_shift: 5.0
    infer_shift: 5.0
    num_train_timesteps: 1000
  action_scheduler:
    train_shift: 5.0
    infer_shift: 5.0
    num_train_timesteps: 1000
  loss:
    lambda_action: 1.0