Datasets:

ArXiv:
License:
kxzxvbk commited on
Commit
09658da
1 Parent(s): 5064de2

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +27 -10
README.md CHANGED
@@ -22,9 +22,16 @@ license: apache-2.0
22
  - [Citation Information](##Citation-Information)
23
  - [Contributions](##Contributions)
24
 
25
- ## Supported Tasks and Leaderboard
26
 
27
- - TODO
 
 
 
 
 
 
 
28
 
29
  ## Dataset Description
30
 
@@ -45,7 +52,7 @@ A data point comprises tuples of sequences of (observations, actions, hidden_sta
45
  ### Data Fields
46
 
47
  - `obs`: An Array3D containing observations from 8 trajectories of an evaluated agent. The data type is uint8 and each value is in 0 to 255.
48
- - `actions`: An integer containing actions from 8 trajectories of an evaluated agent. This value is from 0 to 5.
49
  - `hidden_state`: An Array3D containing corresponding hidden states generated by EfficientZero, from 8 trajectories of an evaluated agent. The data type is float32.
50
 
51
  ### Data Splits
@@ -59,7 +66,7 @@ There is only a training set for this dataset, as evaluation is undertaken by in
59
  ### Source Data
60
  #### Initial Data Collection and Normalization
61
  - This dataset is collected by EfficientZero policy.
62
- - Each return of 8 episodes is 20.
63
  - No normalization is previously applied ( i.e. each value of observation is a uint8 scalar in [0, 255] )
64
  #### Who are the source language producers?
65
  - [@kxzxvbk](https://huggingface.co/kxzxvbk)
@@ -76,20 +83,30 @@ There is only a training set for this dataset, as evaluation is undertaken by in
76
  - This dataset can potentially promote the research for sequence based imitation learning algorithms.
77
 
78
  ### Known Limitations
79
- - TODO
 
80
 
81
  ## Additional Information
82
 
83
- ### Licensing Information
84
- - TODO
85
 
86
  ### Citation Information
87
 
88
  ```
89
- TODO
 
 
 
 
 
 
90
  ```
91
 
92
  ### Contributions
93
- Thanks to [@test](test_url), for adding this dataset.
 
 
 
94
 
95
- [How to contribute to Datasets](https://files.pushshift.io/reddit/)
 
22
  - [Citation Information](##Citation-Information)
23
  - [Contributions](##Contributions)
24
 
25
+ ## Supported Tasks and Baseline
26
 
27
+ - This dataset supports the training for Procedure Cloning algorithm.
28
+
29
+ - Baseline
30
+
31
+ | Length for procedure sequence | Return |
32
+ | ----------------------------- | ------ |
33
+ | 0 | 20 |
34
+ | 4 | -21 |
35
 
36
  ## Dataset Description
37
 
 
52
  ### Data Fields
53
 
54
  - `obs`: An Array3D containing observations from 8 trajectories of an evaluated agent. The data type is uint8 and each value is in 0 to 255.
55
+ - `actions`: An integer containing actions from 8 trajectories of an evaluated agent. This value is from 0 to 5. Details about this environment can be viewed at [Pong - Gym Documentation](https://www.gymlibrary.dev/environments/atari/pong/).
56
  - `hidden_state`: An Array3D containing corresponding hidden states generated by EfficientZero, from 8 trajectories of an evaluated agent. The data type is float32.
57
 
58
  ### Data Splits
 
66
  ### Source Data
67
  #### Initial Data Collection and Normalization
68
  - This dataset is collected by EfficientZero policy.
69
+ - The standard for expert data is that each return of 8 episodes is over 20.
70
  - No normalization is previously applied ( i.e. each value of observation is a uint8 scalar in [0, 255] )
71
  #### Who are the source language producers?
72
  - [@kxzxvbk](https://huggingface.co/kxzxvbk)
 
83
  - This dataset can potentially promote the research for sequence based imitation learning algorithms.
84
 
85
  ### Known Limitations
86
+ - This dataset is only used for academic research.
87
+ - For any commercial use or other cooperation, please contact: opendilab@pjlab.org.cn
88
 
89
  ## Additional Information
90
 
91
+ ### License
92
+ This dataset is under [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0).
93
 
94
  ### Citation Information
95
 
96
  ```
97
+ @misc{Pong-v4-expert-MCTS,
98
+ title={{Pong-v4-expert-MCTS: OpenDILab} A dataset for Procedure Cloning algorithm using Pong-v4.},
99
+ author={Pong-v4-expert-MCTS Contributors},
100
+ publisher = {huggingface},
101
+ howpublished = {\url{https://huggingface.co/datasets/OpenDILabCommunity/Pong-v4-expert-MCTS}},
102
+ year={2023},
103
+ }
104
  ```
105
 
106
  ### Contributions
107
+ This data is partially based on the following repo, many thanks to their pioneering work:
108
+
109
+ - https://github.com/opendilab/DI-engine
110
+ - https://github.com/opendilab/LightZero
111
 
112
+ Please view the [doc](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cardsHow) for anyone who want to contribute to this dataset.