WendiChen commited on
Commit
c4cd568
Β·
verified Β·
1 Parent(s): 8eb991f

[docs] update dataset card

Browse files
Files changed (1) hide show
  1. README.md +32 -5
README.md CHANGED
@@ -14,7 +14,7 @@ task_categories:
14
 
15
  ## Description
16
 
17
- This is the raw dataset used in the paper
18
  [Reactive Diffusion Policy: Slow-Fast Visual-Tactile Policy Learning for Contact-Rich Manipulation](https://reactive-diffusion-policy.github.io).
19
 
20
  - [Paper](https://arxiv.org/pdf/2503.02881)
@@ -26,19 +26,46 @@ This is the raw dataset used in the paper
26
  We offer two versions of the dataset:
27
  one is the [full dataset](https://huggingface.co/datasets/WendiChen/reactive_diffusion_policy_dataset/tree/main/dataset_full) used to train the models in our paper,
28
  and the other is a [mini dataset](https://huggingface.co/datasets/WendiChen/reactive_diffusion_policy_dataset/tree/main/dataset_mini) for easier examination.
29
- Both versions include subsets of peeling, wiping and lifting.
30
- Each subset is structured as follows:
31
 
 
32
  ```
33
  subset_name
34
  β”œβ”€β”€ seq_01.pkl
35
  β”œβ”€β”€ seq_02.pkl
36
  β”œβ”€β”€ ...
37
  ```
38
- Note that we split the full lifting subset into 2 parts due to file size restrictions.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
39
 
40
  ## Usage
41
 
42
  Follow the [README in our GitHub repo](https://github.com/xiaoxiaoxh/reactive_diffusion_policy)
43
- to [postprocess the data](https://github.com/xiaoxiaoxh/reactive_diffusion_policy#data-postprocessing)
44
  and [train the model](https://github.com/xiaoxiaoxh/reactive_diffusion_policy#-training).
 
14
 
15
  ## Description
16
 
17
+ This is the raw and postprocessed dataset used in the paper
18
  [Reactive Diffusion Policy: Slow-Fast Visual-Tactile Policy Learning for Contact-Rich Manipulation](https://reactive-diffusion-policy.github.io).
19
 
20
  - [Paper](https://arxiv.org/pdf/2503.02881)
 
26
  We offer two versions of the dataset:
27
  one is the [full dataset](https://huggingface.co/datasets/WendiChen/reactive_diffusion_policy_dataset/tree/main/dataset_full) used to train the models in our paper,
28
  and the other is a [mini dataset](https://huggingface.co/datasets/WendiChen/reactive_diffusion_policy_dataset/tree/main/dataset_mini) for easier examination.
29
+ Both versions include raw and postprocessed subsets of peeling, wiping and lifting.
 
30
 
31
+ Each raw subset is structured as follows:
32
  ```
33
  subset_name
34
  β”œβ”€β”€ seq_01.pkl
35
  β”œβ”€β”€ seq_02.pkl
36
  β”œβ”€β”€ ...
37
  ```
38
+ > Note that we split the full raw lifting subset into 2 parts due to file size restrictions.
39
+
40
+ Each postprocessed subset is stored in [Zarr](https://zarr.dev) format, which is structured as follows:
41
+ ```
42
+ β”œβ”€β”€ action (25710, 4) float32
43
+ β”œβ”€β”€ external_img (25710, 240, 320, 3) uint8
44
+ β”œβ”€β”€ left_gripper1_img (25710, 240, 320, 3) uint8
45
+ β”œβ”€β”€ left_gripper1_initial_marker (25710, 63, 2) float32
46
+ β”œβ”€β”€ left_gripper1_marker_offset (25710, 63, 2) float32
47
+ β”œβ”€β”€ left_gripper1_marker_offset_emb (25710, 15) float32
48
+ β”œβ”€β”€ left_gripper2_img (25710, 240, 320, 3) uint8
49
+ β”œβ”€β”€ left_gripper2_initial_marker (25710, 25, 2) float32
50
+ β”œβ”€β”€ left_gripper2_marker_offset (25710, 25, 2) float32
51
+ β”œβ”€β”€ left_gripper2_marker_offset_emb (25710, 15) float32
52
+ β”œβ”€β”€ left_robot_gripper_force (25710, 1) float32
53
+ β”œβ”€β”€ left_robot_gripper_width (25710, 1) float32
54
+ β”œβ”€β”€ left_robot_tcp_pose (25710, 9) float32
55
+ β”œβ”€β”€ left_robot_tcp_vel (25710, 6) float32
56
+ β”œβ”€β”€ left_robot_tcp_wrench (25710, 6) float32
57
+ β”œβ”€β”€ left_wrist_img (25710, 240, 320, 3) uint8
58
+ β”œβ”€β”€ right_robot_gripper_force (25710, 1) float32
59
+ β”œβ”€β”€ right_robot_gripper_width (25710, 1) float32
60
+ β”œβ”€β”€ right_robot_tcp_pose (25710, 9) float32
61
+ β”œβ”€β”€ right_robot_tcp_vel (25710, 6) float32
62
+ β”œβ”€β”€ right_robot_tcp_wrench (25710, 6) float32
63
+ β”œβ”€β”€ target (25710, 4) float32
64
+ └── timestamp (25710,) float32
65
+ ```
66
 
67
  ## Usage
68
 
69
  Follow the [README in our GitHub repo](https://github.com/xiaoxiaoxh/reactive_diffusion_policy)
70
+ to [postprocess the raw data](https://github.com/xiaoxiaoxh/reactive_diffusion_policy#data-postprocessing)
71
  and [train the model](https://github.com/xiaoxiaoxh/reactive_diffusion_policy#-training).