the-future-dev commited on
Commit
136e435
·
verified ·
1 Parent(s): f529505

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +36 -3
README.md CHANGED
@@ -11,6 +11,39 @@ configs:
11
  - config_name: default
12
  data_files: data/*/*.parquet
13
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
14
 
15
  This dataset was created using [LeRobot](https://github.com/huggingface/lerobot).
16
 
@@ -29,15 +62,15 @@ This dataset was created using [LeRobot](https://github.com/huggingface/lerobot)
29
  {
30
  "codebase_version": "v2.1",
31
  "robot_type": "2d pointer",
32
- "total_episodes": 1,
33
- "total_frames": 84,
34
  "total_tasks": 1,
35
  "total_videos": 0,
36
  "total_chunks": 1,
37
  "chunks_size": 1000,
38
  "fps": 10,
39
  "splits": {
40
- "train": "0:1"
41
  },
42
  "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
43
  "video_path": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4",
 
11
  - config_name: default
12
  data_files: data/*/*.parquet
13
  ---
14
+ # PushT Dataset
15
+
16
+ This dataset contains demonstrations for the PushT environment, a robotic manipulation task where an agent needs to push a T-shaped object onto a matching target surface.
17
+ Each episode is initialized with a **randomized** T-block and target position and orientation to ensure a more realistic scenario.
18
+
19
+ ## Environment Details
20
+
21
+ The dataset was collected using the [gym-pusht](https://github.com/huggingface/gym-pusht) environment, which provides a simple 2D robotic pushing task.
22
+
23
+ - **Task**: Push the T-shaped gray block onto the T-shaped green target surface using the blue pointer.
24
+ - **Observation Space**: RGB images (96×96×3) and agent position (2D coordinates)
25
+ - **Action Space**: 2D coordinates for the agent position
26
+ - **Success Condition**: The T-shaped block overlaps sufficiently with the target surface
27
+ - **Randomized Goals**: This dataset uses a variation of the environment with `randomize_goal=True`, which randomizes the goal position for each episode, creating a more diverse and challenging dataset.
28
+
29
+ ## Dataset Creation
30
+
31
+ This dataset was created through human demonstrations where a single human operator provided demonstrations of successful task completion.
32
+
33
+ ### Dataset Composition
34
+
35
+ - **States**: RGB images and agent position coordinates
36
+ - **Actions**: 2D position commands for the agent
37
+ - **Rewards**: Sparse rewards based on task completion
38
+ - **Metadata**: Success flags and episode information
39
+
40
+ ## Usage
41
+
42
+ This dataset is intended for training imitation learning and reinforcement learning policies for object manipulation tasks.
43
+
44
+ ### Loading the Dataset
45
+
46
+
47
 
48
  This dataset was created using [LeRobot](https://github.com/huggingface/lerobot).
49
 
 
62
  {
63
  "codebase_version": "v2.1",
64
  "robot_type": "2d pointer",
65
+ "total_episodes": 101,
66
+ "total_frames": 9438,
67
  "total_tasks": 1,
68
  "total_videos": 0,
69
  "total_chunks": 1,
70
  "chunks_size": 1000,
71
  "fps": 10,
72
  "splits": {
73
+ "train": "0:101"
74
  },
75
  "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
76
  "video_path": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4",