solarian-diffusion 0.1.0__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (159) hide show
  1. diffusion_policy/__init__.py +0 -0
  2. diffusion_policy/codecs/__init__.py +0 -0
  3. diffusion_policy/codecs/imagecodecs_numcodecs.py +1408 -0
  4. diffusion_policy/common/__init__.py +0 -0
  5. diffusion_policy/common/checkpoint_util.py +64 -0
  6. diffusion_policy/common/json_logger.py +117 -0
  7. diffusion_policy/common/normalize_util.py +251 -0
  8. diffusion_policy/common/pose_repr_util.py +112 -0
  9. diffusion_policy/common/pytorch_util.py +82 -0
  10. diffusion_policy/common/q_critic_util.py +135 -0
  11. diffusion_policy/common/replay_buffer.py +672 -0
  12. diffusion_policy/common/sampler.py +204 -0
  13. diffusion_policy/config/legacy/task/blockpush_lowdim_seed.yaml +34 -0
  14. diffusion_policy/config/legacy/task/blockpush_lowdim_seed_abs.yaml +34 -0
  15. diffusion_policy/config/legacy/task/can_image.yaml +64 -0
  16. diffusion_policy/config/legacy/task/can_image_abs.yaml +64 -0
  17. diffusion_policy/config/legacy/task/can_lowdim.yaml +45 -0
  18. diffusion_policy/config/legacy/task/can_lowdim_abs.yaml +46 -0
  19. diffusion_policy/config/legacy/task/kitchen_lowdim.yaml +33 -0
  20. diffusion_policy/config/legacy/task/kitchen_lowdim_abs.yaml +38 -0
  21. diffusion_policy/config/legacy/task/lift_abs.yaml +72 -0
  22. diffusion_policy/config/legacy/task/lift_image.yaml +64 -0
  23. diffusion_policy/config/legacy/task/lift_image_abs.yaml +63 -0
  24. diffusion_policy/config/legacy/task/lift_lowdim.yaml +46 -0
  25. diffusion_policy/config/legacy/task/lift_lowdim_abs.yaml +47 -0
  26. diffusion_policy/config/legacy/task/pusht_image.yaml +40 -0
  27. diffusion_policy/config/legacy/task/pusht_lowdim.yaml +34 -0
  28. diffusion_policy/config/legacy/task/real_pusht_image.yaml +47 -0
  29. diffusion_policy/config/legacy/task/square_abs.yaml +72 -0
  30. diffusion_policy/config/legacy/task/square_image.yaml +64 -0
  31. diffusion_policy/config/legacy/task/square_image_abs.yaml +64 -0
  32. diffusion_policy/config/legacy/task/square_lowdim.yaml +46 -0
  33. diffusion_policy/config/legacy/task/square_lowdim_abs.yaml +47 -0
  34. diffusion_policy/config/legacy/task/tool_hang_abs.yaml +71 -0
  35. diffusion_policy/config/legacy/task/tool_hang_image.yaml +63 -0
  36. diffusion_policy/config/legacy/task/tool_hang_image_abs.yaml +63 -0
  37. diffusion_policy/config/legacy/task/tool_hang_lowdim.yaml +45 -0
  38. diffusion_policy/config/legacy/task/tool_hang_lowdim_abs.yaml +46 -0
  39. diffusion_policy/config/legacy/task/transport_image.yaml +75 -0
  40. diffusion_policy/config/legacy/task/transport_image_abs.yaml +75 -0
  41. diffusion_policy/config/legacy/task/transport_lowdim.yaml +49 -0
  42. diffusion_policy/config/legacy/task/transport_lowdim_abs.yaml +49 -0
  43. diffusion_policy/config/legacy/task/umi.yaml +45 -0
  44. diffusion_policy/config/legacy/task/umi_image.yaml +36 -0
  45. diffusion_policy/config/legacy/task/umi_teleop_image.yaml +41 -0
  46. diffusion_policy/config/legacy/train_bet_lowdim_workspace.yaml +132 -0
  47. diffusion_policy/config/legacy/train_diffusion_transformer_hybrid_workspace.yaml +141 -0
  48. diffusion_policy/config/legacy/train_diffusion_transformer_lowdim_kitchen_workspace.yaml +146 -0
  49. diffusion_policy/config/legacy/train_diffusion_transformer_lowdim_pusht_workspace.yaml +146 -0
  50. diffusion_policy/config/legacy/train_diffusion_transformer_lowdim_workspace.yaml +146 -0
  51. diffusion_policy/config/legacy/train_diffusion_transformer_real_hybrid_workspace.yaml +144 -0
  52. diffusion_policy/config/legacy/train_diffusion_transformer_umi_teleop_hybrid_workspace.yaml +156 -0
  53. diffusion_policy/config/legacy/train_diffusion_unet_ddim_hybrid_workspace.yaml +140 -0
  54. diffusion_policy/config/legacy/train_diffusion_unet_ddim_lowdim_workspace.yaml +146 -0
  55. diffusion_policy/config/legacy/train_diffusion_unet_hybrid_workspace.yaml +137 -0
  56. diffusion_policy/config/legacy/train_diffusion_unet_image_pretrained_workspace.yaml +149 -0
  57. diffusion_policy/config/legacy/train_diffusion_unet_image_workspace.yaml +150 -0
  58. diffusion_policy/config/legacy/train_diffusion_unet_lowdim_workspace.yaml +143 -0
  59. diffusion_policy/config/legacy/train_diffusion_unet_real_hybrid_workspace.yaml +141 -0
  60. diffusion_policy/config/legacy/train_diffusion_unet_real_image_workspace.yaml +152 -0
  61. diffusion_policy/config/legacy/train_diffusion_unet_real_pretrained_workspace.yaml +152 -0
  62. diffusion_policy/config/legacy/train_diffusion_unet_timm_workspace.yaml +145 -0
  63. diffusion_policy/config/legacy/train_diffusion_unet_umi_hybrid_workspace.yaml +141 -0
  64. diffusion_policy/config/legacy/train_diffusion_unet_umi_teleop_hybrid_workspace.yaml +153 -0
  65. diffusion_policy/config/legacy/train_diffusion_unet_umi_timm_workspace.yaml +145 -0
  66. diffusion_policy/config/legacy/train_diffusion_unet_umi_workspace.yaml +150 -0
  67. diffusion_policy/config/legacy/train_diffusion_unet_video_workspace.yaml +128 -0
  68. diffusion_policy/config/legacy/train_ibc_dfo_hybrid_workspace.yaml +111 -0
  69. diffusion_policy/config/legacy/train_ibc_dfo_lowdim_workspace.yaml +109 -0
  70. diffusion_policy/config/legacy/train_ibc_dfo_real_hybrid_workspace.yaml +111 -0
  71. diffusion_policy/config/legacy/train_robomimic_image_workspace.yaml +92 -0
  72. diffusion_policy/config/legacy/train_robomimic_lowdim_workspace.yaml +92 -0
  73. diffusion_policy/config/legacy/train_robomimic_real_image_workspace.yaml +92 -0
  74. diffusion_policy/config/rl_annotate.yaml +265 -0
  75. diffusion_policy/config/rl_annotate_per_frame.yaml +162 -0
  76. diffusion_policy/config/rl_annotate_per_frame_v4.yaml +162 -0
  77. diffusion_policy/config/rl_annotate_v2.yaml +141 -0
  78. diffusion_policy/config/rl_annotate_v3.yaml +165 -0
  79. diffusion_policy/config/task/lift.yaml +89 -0
  80. diffusion_policy/config/task/square.yaml +92 -0
  81. diffusion_policy/config/task/tool_hang.yaml +93 -0
  82. diffusion_policy/config/task/tool_hang_abs.yaml +71 -0
  83. diffusion_policy/config/task/umi.yaml +105 -0
  84. diffusion_policy/config/task/umi_bimanual.yaml +135 -0
  85. diffusion_policy/config/task/umi_cup_wild.yaml +106 -0
  86. diffusion_policy/config/task/umi_image.yaml +36 -0
  87. diffusion_policy/config/task/umi_pick_lego.yaml +106 -0
  88. diffusion_policy/config/task/umi_pick_single_lego_175.yaml +106 -0
  89. diffusion_policy/config/task/umi_pick_single_lego_bbox.yaml +135 -0
  90. diffusion_policy/config/task/umi_teleop_image.yaml +41 -0
  91. diffusion_policy/config/train_diffusion_transformer_umi_bimanual_workspace.yaml +185 -0
  92. diffusion_policy/config/train_diffusion_transformer_umi_workspace.yaml +185 -0
  93. diffusion_policy/config/train_diffusion_unet_image_workspace.yaml +152 -0
  94. diffusion_policy/config/train_diffusion_unet_timm_robomimic_workspace.yaml +152 -0
  95. diffusion_policy/config/train_diffusion_unet_timm_umi_cup_wild.yaml +260 -0
  96. diffusion_policy/config/train_diffusion_unet_timm_umi_pick_lego_pile_bbox.yaml +265 -0
  97. diffusion_policy/config/train_diffusion_unet_timm_umi_pick_lego_pile_bbox_awr.yaml +324 -0
  98. diffusion_policy/config/train_diffusion_unet_timm_umi_pick_lego_pile_bbox_parl.yaml +399 -0
  99. diffusion_policy/config/train_diffusion_unet_timm_umi_pick_single_lego.yaml +258 -0
  100. diffusion_policy/config/train_diffusion_unet_timm_umi_pick_single_lego_125.yaml +260 -0
  101. diffusion_policy/config/train_diffusion_unet_timm_umi_pick_single_lego_175.yaml +260 -0
  102. diffusion_policy/config/train_diffusion_unet_timm_umi_pick_single_lego_bbox.yaml +265 -0
  103. diffusion_policy/config/train_diffusion_unet_timm_umi_pick_two_legos_bbox.yaml +265 -0
  104. diffusion_policy/config/train_diffusion_unet_timm_umi_workspace.yaml +224 -0
  105. diffusion_policy/config/train_diffusion_unet_timm_umi_workspace_lego.yaml +213 -0
  106. diffusion_policy/config/train_diffusion_unet_timm_umi_workspace_lego_resume.yaml +213 -0
  107. diffusion_policy/config/train_diffusion_unet_timm_umi_workspace_pick_lego.yaml +254 -0
  108. diffusion_policy/config/train_diffusion_unet_timm_umi_workspace_red_cup.yaml +219 -0
  109. diffusion_policy/config/train_diffusion_unet_umi_bimanual_workspace.yaml +203 -0
  110. diffusion_policy/config/train_iql_umi.yaml +186 -0
  111. diffusion_policy/config/train_iql_umi_v2.yaml +212 -0
  112. diffusion_policy/config/train_iql_umi_v3.yaml +212 -0
  113. diffusion_policy/dataset/__init__.py +0 -0
  114. diffusion_policy/dataset/base_dataset.py +76 -0
  115. diffusion_policy/dataset/umi_dataset.py +652 -0
  116. diffusion_policy/env_runner/__init__.py +0 -0
  117. diffusion_policy/env_runner/base_image_runner.py +9 -0
  118. diffusion_policy/env_runner/real_pusht_image_runner.py +10 -0
  119. diffusion_policy/model/__init__.py +0 -0
  120. diffusion_policy/model/common/__init__.py +0 -0
  121. diffusion_policy/model/common/dict_of_tensor_mixin.py +38 -0
  122. diffusion_policy/model/common/lr_scheduler.py +46 -0
  123. diffusion_policy/model/common/module_attr_mixin.py +14 -0
  124. diffusion_policy/model/common/normalizer.py +353 -0
  125. diffusion_policy/model/common/tensor_util.py +960 -0
  126. diffusion_policy/model/diffusion/__init__.py +0 -0
  127. diffusion_policy/model/diffusion/conditional_unet1d.py +238 -0
  128. diffusion_policy/model/diffusion/conv1d_components.py +46 -0
  129. diffusion_policy/model/diffusion/ema_model.py +88 -0
  130. diffusion_policy/model/diffusion/mask_generator.py +221 -0
  131. diffusion_policy/model/diffusion/positional_embedding.py +17 -0
  132. diffusion_policy/model/vision/__init__.py +0 -0
  133. diffusion_policy/model/vision/crop_randomizer.py +288 -0
  134. diffusion_policy/model/vision/multi_image_obs_encoder.py +195 -0
  135. diffusion_policy/model/vision/timm_obs_encoder.py +525 -0
  136. diffusion_policy/policy/__init__.py +0 -0
  137. diffusion_policy/policy/base_image_policy.py +27 -0
  138. diffusion_policy/policy/diffusion_unet_image_policy.py +264 -0
  139. diffusion_policy/policy/diffusion_unet_timm_policy.py +607 -0
  140. diffusion_policy/policy/parl_action_optimizer.py +675 -0
  141. diffusion_policy/workspace/__init__.py +0 -0
  142. diffusion_policy/workspace/base_workspace.py +146 -0
  143. diffusion_policy/workspace/train_diffusion_unet_image_workspace.py +1585 -0
  144. solarian_diffusion/__init__.py +32 -0
  145. solarian_diffusion/config.py +31 -0
  146. solarian_diffusion/policy_server.py +316 -0
  147. solarian_diffusion/rtc.py +272 -0
  148. solarian_diffusion/service.py +16 -0
  149. solarian_diffusion/trainer.py +213 -0
  150. solarian_diffusion-0.1.0.dist-info/METADATA +126 -0
  151. solarian_diffusion-0.1.0.dist-info/RECORD +159 -0
  152. solarian_diffusion-0.1.0.dist-info/WHEEL +5 -0
  153. solarian_diffusion-0.1.0.dist-info/entry_points.txt +8 -0
  154. solarian_diffusion-0.1.0.dist-info/licenses/LICENSE +201 -0
  155. solarian_diffusion-0.1.0.dist-info/licenses/NOTICE +45 -0
  156. solarian_diffusion-0.1.0.dist-info/top_level.txt +3 -0
  157. umi/__init__.py +0 -0
  158. umi/common/__init__.py +0 -0
  159. umi/common/pose_util.py +115 -0
@@ -0,0 +1,204 @@
1
+ from typing import Optional
2
+ import numpy as np
3
+ import random
4
+ import scipy.interpolate as si
5
+ import scipy.spatial.transform as st
6
+ from diffusion_policy.common.replay_buffer import ReplayBuffer
7
+
8
+ def get_val_mask(n_episodes, val_ratio, seed=0):
9
+ val_mask = np.zeros(n_episodes, dtype=bool)
10
+ if val_ratio <= 0:
11
+ return val_mask
12
+
13
+ # have at least 1 episode for validation, and at least 1 episode for train
14
+ n_val = min(max(1, round(n_episodes * val_ratio)), n_episodes-1)
15
+ rng = np.random.default_rng(seed=seed)
16
+ val_idxs = rng.choice(n_episodes, size=n_val, replace=False)
17
+ val_mask[val_idxs] = True
18
+ return val_mask
19
+
20
+
21
+ class SequenceSampler:
22
+ def __init__(self,
23
+ shape_meta: dict,
24
+ replay_buffer: ReplayBuffer,
25
+ rgb_keys: list,
26
+ lowdim_keys: list,
27
+ key_horizon: dict,
28
+ key_latency_steps: dict,
29
+ key_down_sample_steps: dict,
30
+ episode_mask: Optional[np.ndarray]=None,
31
+ action_padding: bool=False,
32
+ repeat_frame_prob: float=0.0,
33
+ max_duration: Optional[float]=None
34
+ ):
35
+ episode_ends = replay_buffer.episode_ends[:]
36
+
37
+ # load gripper_width
38
+ gripper_width = replay_buffer['robot0_gripper_width'][:, 0]
39
+ gripper_width_threshold = 0.08
40
+ self.repeat_frame_prob = repeat_frame_prob
41
+
42
+ # create indices, including (current_idx, start_idx, end_idx)
43
+ indices = list()
44
+ for i in range(len(episode_ends)):
45
+ before_first_grasp = True # initialize for each episode
46
+ if episode_mask is not None and not episode_mask[i]:
47
+ # skip episode
48
+ continue
49
+ start_idx = 0 if i == 0 else episode_ends[i-1]
50
+ end_idx = episode_ends[i]
51
+ if max_duration is not None:
52
+ end_idx = min(end_idx, max_duration * 60)
53
+ for current_idx in range(start_idx, end_idx):
54
+ if not action_padding and end_idx < current_idx + (key_horizon['action'] - 1) * key_down_sample_steps['action'] + 1:
55
+ continue
56
+ if gripper_width[current_idx] < gripper_width_threshold:
57
+ before_first_grasp = False
58
+ indices.append((current_idx, start_idx, end_idx, before_first_grasp))
59
+
60
+ # load low_dim to memory and keep rgb as compressed zarr array
61
+ self.replay_buffer = dict()
62
+ self.num_robot = 0
63
+ for key in lowdim_keys:
64
+ if key.endswith('eef_pos'):
65
+ self.num_robot += 1
66
+
67
+ if key.endswith('pos_abs'):
68
+ axis = shape_meta['obs'][key]['axis']
69
+ if isinstance(axis, int):
70
+ axis = [axis]
71
+ self.replay_buffer[key] = replay_buffer[key[:-4]][:, list(axis)]
72
+ elif key.endswith('quat_abs'):
73
+ axis = shape_meta['obs'][key]['axis']
74
+ if isinstance(axis, int):
75
+ axis = [axis]
76
+ # HACK for hybrid abs/relative proprioception
77
+ rot_in = replay_buffer[key[:-4]][:]
78
+ rot_out = st.Rotation.from_quat(rot_in).as_euler('XYZ')
79
+ self.replay_buffer[key] = rot_out[:, list(axis)]
80
+ elif key.endswith('axis_angle_abs'):
81
+ axis = shape_meta['obs'][key]['axis']
82
+ if isinstance(axis, int):
83
+ axis = [axis]
84
+ rot_in = replay_buffer[key[:-4]][:]
85
+ rot_out = st.Rotation.from_rotvec(rot_in).as_euler('XYZ')
86
+ self.replay_buffer[key] = rot_out[:, list(axis)]
87
+ else:
88
+ self.replay_buffer[key] = replay_buffer[key][:]
89
+ for key in rgb_keys:
90
+ self.replay_buffer[key] = replay_buffer[key]
91
+
92
+
93
+ if 'action' in replay_buffer:
94
+ self.replay_buffer['action'] = replay_buffer['action'][:]
95
+ else:
96
+ # construct action (concatenation of [eef_pos, eef_rot, gripper_width])
97
+ actions = list()
98
+ for robot_idx in range(self.num_robot):
99
+ for cat in ['eef_pos', 'eef_rot_axis_angle', 'gripper_width']:
100
+ key = f'robot{robot_idx}_{cat}'
101
+ if key in self.replay_buffer:
102
+ actions.append(self.replay_buffer[key])
103
+ self.replay_buffer['action'] = np.concatenate(actions, axis=-1)
104
+
105
+ self.action_padding = action_padding
106
+ self.indices = indices
107
+ self.rgb_keys = rgb_keys
108
+ self.lowdim_keys = lowdim_keys
109
+ self.key_horizon = key_horizon
110
+ self.key_latency_steps = key_latency_steps
111
+ self.key_down_sample_steps = key_down_sample_steps
112
+
113
+ self.ignore_rgb_is_applied = False # speed up the interation when getting normalizaer
114
+
115
+ def __len__(self):
116
+ return len(self.indices)
117
+
118
+ def sample_sequence(self, idx):
119
+ current_idx, start_idx, end_idx, before_first_grasp = self.indices[idx]
120
+
121
+ result = dict()
122
+
123
+ obs_keys = self.rgb_keys + self.lowdim_keys
124
+ if self.ignore_rgb_is_applied:
125
+ obs_keys = self.lowdim_keys
126
+
127
+ # observation
128
+ for key in obs_keys:
129
+ input_arr = self.replay_buffer[key]
130
+ this_horizon = self.key_horizon[key]
131
+ this_latency_steps = self.key_latency_steps[key]
132
+ this_downsample_steps = self.key_down_sample_steps[key]
133
+
134
+ if key in self.rgb_keys:
135
+ assert this_latency_steps == 0
136
+ num_valid = min(this_horizon, (current_idx - start_idx) // this_downsample_steps + 1)
137
+ slice_start = current_idx - (num_valid - 1) * this_downsample_steps
138
+
139
+ output = input_arr[slice_start: current_idx + 1: this_downsample_steps]
140
+ assert output.shape[0] == num_valid
141
+
142
+ # solve padding
143
+ if output.shape[0] < this_horizon:
144
+ padding = np.repeat(output[:1], this_horizon - output.shape[0], axis=0)
145
+ output = np.concatenate([padding, output], axis=0)
146
+ else:
147
+ idx_with_latency = np.array(
148
+ [current_idx - idx * this_downsample_steps + this_latency_steps for idx in range(this_horizon)],
149
+ dtype=np.float32)
150
+ idx_with_latency = idx_with_latency[::-1]
151
+ idx_with_latency = np.clip(idx_with_latency, start_idx, end_idx - 1)
152
+ interpolation_start = max(int(idx_with_latency[0]) - 5, start_idx)
153
+ interpolation_end = min(int(idx_with_latency[-1]) + 2 + 5, end_idx)
154
+
155
+ if 'rot' in key:
156
+ # rotation
157
+ rot_preprocess, rot_postprocess = None, None
158
+ if key.endswith('quat'):
159
+ rot_preprocess = st.Rotation.from_quat
160
+ rot_postprocess = st.Rotation.as_quat
161
+ elif key.endswith('axis_angle'):
162
+ rot_preprocess = st.Rotation.from_rotvec
163
+ rot_postprocess = st.Rotation.as_rotvec
164
+ else:
165
+ raise NotImplementedError
166
+ slerp = st.Slerp(
167
+ times=np.arange(interpolation_start, interpolation_end),
168
+ rotations=rot_preprocess(input_arr[interpolation_start: interpolation_end]))
169
+ output = rot_postprocess(slerp(idx_with_latency))
170
+ else:
171
+ interp = si.interp1d(
172
+ x=np.arange(interpolation_start, interpolation_end),
173
+ y=input_arr[interpolation_start: interpolation_end],
174
+ axis=0, assume_sorted=True)
175
+ output = interp(idx_with_latency)
176
+
177
+ result[key] = output
178
+
179
+ # repeat frame before first grasp
180
+ if self.repeat_frame_prob != 0.0:
181
+ if before_first_grasp and random.random() < self.repeat_frame_prob:
182
+ for key in obs_keys:
183
+ result[key][:-1] = result[key][-1:]
184
+
185
+ # aciton
186
+ input_arr = self.replay_buffer['action']
187
+ action_horizon = self.key_horizon['action']
188
+ action_latency_steps = self.key_latency_steps['action']
189
+ assert action_latency_steps == 0
190
+ action_down_sample_steps = self.key_down_sample_steps['action']
191
+ slice_end = min(end_idx, current_idx + (action_horizon - 1) * action_down_sample_steps + 1)
192
+ output = input_arr[current_idx: slice_end: action_down_sample_steps]
193
+ # solve padding
194
+ if not self.action_padding:
195
+ assert output.shape[0] == action_horizon
196
+ elif output.shape[0] < action_horizon:
197
+ padding = np.repeat(output[-1:], action_horizon - output.shape[0], axis=0)
198
+ output = np.concatenate([output, padding], axis=0)
199
+ result['action'] = output
200
+
201
+ return result
202
+
203
+ def ignore_rgb(self, apply=True):
204
+ self.ignore_rgb_is_applied = apply
@@ -0,0 +1,34 @@
1
+ name: blockpush_lowdim_seed
2
+
3
+ obs_dim: 16
4
+ action_dim: 2
5
+ keypoint_dim: 2
6
+ obs_eef_target: True
7
+
8
+ env_runner:
9
+ _target_: diffusion_policy.env_runner.blockpush_lowdim_runner.BlockPushLowdimRunner
10
+ n_train: 6
11
+ n_train_vis: 2
12
+ train_start_seed: 0
13
+ n_test: 50
14
+ n_test_vis: 4
15
+ test_start_seed: 100000
16
+ max_steps: 350
17
+ n_obs_steps: ${n_obs_steps}
18
+ n_action_steps: ${n_action_steps}
19
+ fps: 5
20
+ past_action: ${past_action_visible}
21
+ abs_action: False
22
+ obs_eef_target: ${task.obs_eef_target}
23
+ n_envs: null
24
+
25
+ dataset:
26
+ _target_: diffusion_policy.dataset.blockpush_lowdim_dataset.BlockPushLowdimDataset
27
+ zarr_path: data/block_pushing/multimodal_push_seed.zarr
28
+ horizon: ${horizon}
29
+ pad_before: ${eval:'${n_obs_steps}-1'}
30
+ pad_after: ${eval:'${horizon}-${n_obs_steps}'}
31
+ obs_eef_target: ${task.obs_eef_target}
32
+ use_manual_normalizer: False
33
+ seed: 42
34
+ val_ratio: 0.02
@@ -0,0 +1,34 @@
1
+ name: blockpush_lowdim_seed_abs
2
+
3
+ obs_dim: 16
4
+ action_dim: 2
5
+ keypoint_dim: 2
6
+ obs_eef_target: True
7
+
8
+ env_runner:
9
+ _target_: diffusion_policy.env_runner.blockpush_lowdim_runner.BlockPushLowdimRunner
10
+ n_train: 6
11
+ n_train_vis: 2
12
+ train_start_seed: 0
13
+ n_test: 50
14
+ n_test_vis: 4
15
+ test_start_seed: 100000
16
+ max_steps: 350
17
+ n_obs_steps: ${n_obs_steps}
18
+ n_action_steps: ${n_action_steps}
19
+ fps: 5
20
+ past_action: ${past_action_visible}
21
+ abs_action: True
22
+ obs_eef_target: ${task.obs_eef_target}
23
+ n_envs: null
24
+
25
+ dataset:
26
+ _target_: diffusion_policy.dataset.blockpush_lowdim_dataset.BlockPushLowdimDataset
27
+ zarr_path: data/block_pushing/multimodal_push_seed_abs.zarr
28
+ horizon: ${horizon}
29
+ pad_before: ${eval:'${n_obs_steps}-1'}
30
+ pad_after: ${eval:'${horizon}-${n_obs_steps}'}
31
+ obs_eef_target: ${task.obs_eef_target}
32
+ use_manual_normalizer: False
33
+ seed: 42
34
+ val_ratio: 0.02
@@ -0,0 +1,64 @@
1
+ name: can_image
2
+
3
+ shape_meta: &shape_meta
4
+ # acceptable types: rgb, low_dim
5
+ obs:
6
+ agentview_image:
7
+ shape: [3, 84, 84]
8
+ type: rgb
9
+ robot0_eye_in_hand_image:
10
+ shape: [3, 84, 84]
11
+ type: rgb
12
+ robot0_eef_pos:
13
+ shape: [3]
14
+ # type default: low_dim
15
+ robot0_eef_quat:
16
+ shape: [4]
17
+ robot0_gripper_qpos:
18
+ shape: [2]
19
+ action:
20
+ shape: [7]
21
+
22
+ task_name: &task_name can
23
+ dataset_type: &dataset_type ph
24
+ dataset_path: &dataset_path data/robomimic/datasets/${task.task_name}/${task.dataset_type}/image.hdf5
25
+ abs_action: &abs_action False
26
+
27
+ env_runner:
28
+ _target_: diffusion_policy.env_runner.robomimic_image_runner.RobomimicImageRunner
29
+ dataset_path: *dataset_path
30
+ shape_meta: *shape_meta
31
+ # costs 1GB per env
32
+ n_train: 6
33
+ n_train_vis: 2
34
+ train_start_idx: 0
35
+ n_test: 50
36
+ n_test_vis: 4
37
+ test_start_seed: 100000
38
+ # use python's eval function as resolver, single-quoted string as argument
39
+ max_steps: ${eval:'500 if "${task.dataset_type}" == "mh" else 400'}
40
+ n_obs_steps: ${n_obs_steps}
41
+ n_action_steps: ${n_action_steps}
42
+ render_obs_key: 'agentview_image'
43
+ fps: 10
44
+ crf: 22
45
+ past_action: ${past_action_visible}
46
+ abs_action: *abs_action
47
+ tqdm_interval_sec: 1.0
48
+ n_envs: 28
49
+ # evaluation at this config requires a 16 core 64GB instance.
50
+
51
+ dataset:
52
+ _target_: diffusion_policy.dataset.robomimic_replay_image_dataset.RobomimicReplayImageDataset
53
+ shape_meta: *shape_meta
54
+ dataset_path: *dataset_path
55
+ horizon: ${horizon}
56
+ pad_before: ${eval:'${n_obs_steps}-1+${n_latency_steps}'}
57
+ pad_after: ${eval:'${horizon}-${n_obs_steps}'}
58
+ n_obs_steps: ${dataset_obs_steps}
59
+ abs_action: *abs_action
60
+ rotation_rep: 'rotation_6d'
61
+ use_legacy_normalizer: False
62
+ use_cache: True
63
+ seed: 42
64
+ val_ratio: 0.02
@@ -0,0 +1,64 @@
1
+ name: can_image
2
+
3
+ shape_meta: &shape_meta
4
+ # acceptable types: rgb, low_dim
5
+ obs:
6
+ agentview_image:
7
+ shape: [3, 84, 84]
8
+ type: rgb
9
+ robot0_eye_in_hand_image:
10
+ shape: [3, 84, 84]
11
+ type: rgb
12
+ robot0_eef_pos:
13
+ shape: [3]
14
+ # type default: low_dim
15
+ robot0_eef_quat:
16
+ shape: [4]
17
+ robot0_gripper_qpos:
18
+ shape: [2]
19
+ action:
20
+ shape: [10]
21
+
22
+ task_name: &task_name can
23
+ dataset_type: &dataset_type ph
24
+ dataset_path: &dataset_path data/robomimic/datasets/${task.task_name}/${task.dataset_type}/image_abs.hdf5
25
+ abs_action: &abs_action True
26
+
27
+ env_runner:
28
+ _target_: diffusion_policy.env_runner.robomimic_image_runner.RobomimicImageRunner
29
+ dataset_path: *dataset_path
30
+ shape_meta: *shape_meta
31
+ # costs 1GB per env
32
+ n_train: 6
33
+ n_train_vis: 2
34
+ train_start_idx: 0
35
+ n_test: 50
36
+ n_test_vis: 4
37
+ test_start_seed: 100000
38
+ # use python's eval function as resolver, single-quoted string as argument
39
+ max_steps: ${eval:'500 if "${task.dataset_type}" == "mh" else 400'}
40
+ n_obs_steps: ${n_obs_steps}
41
+ n_action_steps: ${n_action_steps}
42
+ render_obs_key: 'agentview_image'
43
+ fps: 10
44
+ crf: 22
45
+ past_action: ${past_action_visible}
46
+ abs_action: *abs_action
47
+ tqdm_interval_sec: 1.0
48
+ n_envs: 28
49
+ # evaluation at this config requires a 16 core 64GB instance.
50
+
51
+ dataset:
52
+ _target_: diffusion_policy.dataset.robomimic_replay_image_dataset.RobomimicReplayImageDataset
53
+ shape_meta: *shape_meta
54
+ dataset_path: *dataset_path
55
+ horizon: ${horizon}
56
+ pad_before: ${eval:'${n_obs_steps}-1+${n_latency_steps}'}
57
+ pad_after: ${eval:'${horizon}-${n_obs_steps}'}
58
+ n_obs_steps: ${dataset_obs_steps}
59
+ abs_action: *abs_action
60
+ rotation_rep: 'rotation_6d'
61
+ use_legacy_normalizer: False
62
+ use_cache: True
63
+ seed: 42
64
+ val_ratio: 0.02
@@ -0,0 +1,45 @@
1
+ name: can_lowdim
2
+
3
+ obs_dim: 23
4
+ action_dim: 7
5
+ keypoint_dim: 3
6
+
7
+ obs_keys: &obs_keys ['object', 'robot0_eef_pos', 'robot0_eef_quat', 'robot0_gripper_qpos']
8
+ task_name: &task_name can
9
+ dataset_type: &dataset_type ph
10
+ dataset_path: &dataset_path data/robomimic/datasets/${task.task_name}/${task.dataset_type}/low_dim.hdf5
11
+ abs_action: &abs_action False
12
+
13
+ env_runner:
14
+ _target_: diffusion_policy.env_runner.robomimic_lowdim_runner.RobomimicLowdimRunner
15
+ dataset_path: *dataset_path
16
+ obs_keys: *obs_keys
17
+ n_train: 6
18
+ n_train_vis: 2
19
+ train_start_idx: 0
20
+ n_test: 50
21
+ n_test_vis: 4
22
+ test_start_seed: 100000
23
+ # use python's eval function as resolver, single-quoted string as argument
24
+ max_steps: ${eval:'500 if "${task.dataset_type}" == "mh" else 400'}
25
+ n_obs_steps: ${n_obs_steps}
26
+ n_action_steps: ${n_action_steps}
27
+ n_latency_steps: ${n_latency_steps}
28
+ render_hw: [128,128]
29
+ fps: 10
30
+ crf: 22
31
+ past_action: ${past_action_visible}
32
+ abs_action: *abs_action
33
+ n_envs: 28
34
+
35
+ dataset:
36
+ _target_: diffusion_policy.dataset.robomimic_replay_lowdim_dataset.RobomimicReplayLowdimDataset
37
+ dataset_path: *dataset_path
38
+ horizon: ${horizon}
39
+ pad_before: ${eval:'${n_obs_steps}-1+${n_latency_steps}'}
40
+ pad_after: ${eval:'${horizon}-${n_obs_steps}'}
41
+ obs_keys: *obs_keys
42
+ abs_action: *abs_action
43
+ use_legacy_normalizer: False
44
+ seed: 42
45
+ val_ratio: 0.02
@@ -0,0 +1,46 @@
1
+ name: can_lowdim
2
+
3
+ obs_dim: 23
4
+ action_dim: 10
5
+ keypoint_dim: 3
6
+
7
+ obs_keys: &obs_keys ['object', 'robot0_eef_pos', 'robot0_eef_quat', 'robot0_gripper_qpos']
8
+ task_name: &task_name can
9
+ dataset_type: &dataset_type ph
10
+ dataset_path: &dataset_path data/robomimic/datasets/${task.task_name}/${task.dataset_type}/low_dim_abs.hdf5
11
+ abs_action: &abs_action True
12
+
13
+ env_runner:
14
+ _target_: diffusion_policy.env_runner.robomimic_lowdim_runner.RobomimicLowdimRunner
15
+ dataset_path: *dataset_path
16
+ obs_keys: *obs_keys
17
+ n_train: 6
18
+ n_train_vis: 2
19
+ train_start_idx: 0
20
+ n_test: 50
21
+ n_test_vis: 4
22
+ test_start_seed: 100000
23
+ # use python's eval function as resolver, single-quoted string as argument
24
+ max_steps: ${eval:'500 if "${task.dataset_type}" == "mh" else 400'}
25
+ n_obs_steps: ${n_obs_steps}
26
+ n_action_steps: ${n_action_steps}
27
+ n_latency_steps: ${n_latency_steps}
28
+ render_hw: [128,128]
29
+ fps: 10
30
+ crf: 22
31
+ past_action: ${past_action_visible}
32
+ abs_action: *abs_action
33
+ n_envs: 28
34
+
35
+ dataset:
36
+ _target_: diffusion_policy.dataset.robomimic_replay_lowdim_dataset.RobomimicReplayLowdimDataset
37
+ dataset_path: *dataset_path
38
+ horizon: ${horizon}
39
+ pad_before: ${eval:'${n_obs_steps}-1+${n_latency_steps}'}
40
+ pad_after: ${eval:'${horizon}-${n_obs_steps}'}
41
+ obs_keys: *obs_keys
42
+ abs_action: *abs_action
43
+ use_legacy_normalizer: False
44
+ rotation_rep: rotation_6d
45
+ seed: 42
46
+ val_ratio: 0.02
@@ -0,0 +1,33 @@
1
+ name: kitchen_lowdim
2
+
3
+ obs_dim: 60
4
+ action_dim: 9
5
+ keypoint_dim: 3
6
+
7
+ dataset_dir: &dataset_dir data/kitchen
8
+
9
+ env_runner:
10
+ _target_: diffusion_policy.env_runner.kitchen_lowdim_runner.KitchenLowdimRunner
11
+ dataset_dir: *dataset_dir
12
+ n_train: 6
13
+ n_train_vis: 2
14
+ train_start_seed: 0
15
+ n_test: 50
16
+ n_test_vis: 4
17
+ test_start_seed: 100000
18
+ max_steps: 280
19
+ n_obs_steps: ${n_obs_steps}
20
+ n_action_steps: ${n_action_steps}
21
+ render_hw: [240, 360]
22
+ fps: 12.5
23
+ past_action: ${past_action_visible}
24
+ n_envs: null
25
+
26
+ dataset:
27
+ _target_: diffusion_policy.dataset.kitchen_lowdim_dataset.KitchenLowdimDataset
28
+ dataset_dir: *dataset_dir
29
+ horizon: ${horizon}
30
+ pad_before: ${eval:'${n_obs_steps}-1'}
31
+ pad_after: ${eval:'${horizon}-${n_obs_steps}'}
32
+ seed: 42
33
+ val_ratio: 0.02
@@ -0,0 +1,38 @@
1
+ name: kitchen_lowdim
2
+
3
+ obs_dim: 60
4
+ action_dim: 9
5
+ keypoint_dim: 3
6
+
7
+ abs_action: True
8
+ robot_noise_ratio: 0.1
9
+
10
+ env_runner:
11
+ _target_: diffusion_policy.env_runner.kitchen_lowdim_runner.KitchenLowdimRunner
12
+ dataset_dir: data/kitchen
13
+ n_train: 6
14
+ n_train_vis: 2
15
+ train_start_seed: 0
16
+ n_test: 50
17
+ n_test_vis: 4
18
+ test_start_seed: 100000
19
+ max_steps: 280
20
+ n_obs_steps: ${n_obs_steps}
21
+ n_action_steps: ${n_action_steps}
22
+ render_hw: [240, 360]
23
+ fps: 12.5
24
+ past_action: ${past_action_visible}
25
+ abs_action: ${task.abs_action}
26
+ robot_noise_ratio: ${task.robot_noise_ratio}
27
+ n_envs: null
28
+
29
+ dataset:
30
+ _target_: diffusion_policy.dataset.kitchen_mjl_lowdim_dataset.KitchenMjlLowdimDataset
31
+ dataset_dir: data/kitchen/kitchen_demos_multitask
32
+ horizon: ${horizon}
33
+ pad_before: ${eval:'${n_obs_steps}-1'}
34
+ pad_after: ${eval:'${horizon}-${n_obs_steps}'}
35
+ abs_action: ${task.abs_action}
36
+ robot_noise_ratio: ${task.robot_noise_ratio}
37
+ seed: 42
38
+ val_ratio: 0.02
@@ -0,0 +1,72 @@
1
+ name: lift_image
2
+
3
+ shape_meta: &shape_meta
4
+ # acceptable types: rgb, low_dim
5
+ obs:
6
+ agentview_image:
7
+ shape: [3, 84, 84]
8
+ horizon: 2
9
+ type: rgb
10
+ robot0_eye_in_hand_image:
11
+ shape: [3, 84, 84]
12
+ horizon: 2
13
+ type: rgb
14
+ robot0_eef_pos:
15
+ shape: [3]
16
+ horizon: 2
17
+ type: low_dim
18
+ robot0_eef_quat:
19
+ raw_shape: [4]
20
+ shape: [6]
21
+ horizon: 2
22
+ type: low_dim
23
+ rotation_rep: rotation_6d
24
+ robot0_gripper_qpos:
25
+ shape: [2]
26
+ horizon: 2
27
+ type: low_dim
28
+ action:
29
+ shape: [10]
30
+ rotation_rep: rotation_6d
31
+
32
+ task_name: &task_name lift
33
+ dataset_type: &dataset_type ph
34
+ dataset_path: &dataset_path data/robomimic/datasets/${task.task_name}/${task.dataset_type}/image_abs.hdf5
35
+ pose_repr: &pose_repr
36
+ obs_pose_repr: rel # abs or rel
37
+ action_pose_repr: rel
38
+
39
+ env_runner:
40
+ _target_: diffusion_policy.env_runner.robomimic_runner.RobomimicRunner
41
+ dataset_path: *dataset_path
42
+ pose_repr: *pose_repr
43
+ shape_meta: *shape_meta
44
+ n_train: 6
45
+ n_train_vis: 2
46
+ train_start_idx: 0
47
+ n_test: 50
48
+ n_test_vis: 4
49
+ test_start_seed: 100000
50
+ # use python's eval function as resolver, single-quoted string as argument
51
+ max_steps: ${eval:'500 if "${task.dataset_type}" == "mh" else 400'}
52
+ n_obs_steps: ${n_obs_steps}
53
+ n_action_steps: ${n_action_steps}
54
+ render_obs_key: 'agentview_image'
55
+ fps: 10
56
+ crf: 22
57
+ tqdm_interval_sec: 1.0
58
+ n_envs: 28
59
+ # evaluation at this config requires a 16 core 64GB instance.
60
+
61
+ dataset:
62
+ _target_: diffusion_policy.dataset.robomimic_replay_dataset.RobomimicReplayDataset
63
+ pose_repr: *pose_repr
64
+ shape_meta: *shape_meta
65
+ dataset_path: *dataset_path
66
+ horizon: ${horizon}
67
+ pad_before: ${eval:'${n_obs_steps}-1+${n_latency_steps}'}
68
+ pad_after: ${eval:'${horizon}-${n_obs_steps}'}
69
+ n_obs_steps: ${dataset_obs_steps}
70
+ use_cache: True
71
+ seed: 42
72
+ val_ratio: 0.02