Dataset Viewer
Auto-converted to Parquet Duplicate
Search is not available for this dataset
image
image
episode
int32
frame_index
int32
action
int32
1
0
2
1
1
2
1
2
11
1
3
11
1
4
11
1
5
11
1
6
17
1
7
17
1
8
17
1
9
17
1
10
4
1
11
4
1
12
4
1
13
4
1
14
8
1
15
8
1
16
8
1
17
8
1
18
8
1
19
8
1
20
8
1
21
8
1
22
7
1
23
7
1
24
7
1
25
7
1
26
2
1
27
2
1
28
2
1
29
2
1
30
2
1
31
2
1
32
2
1
33
2
1
34
7
1
35
7
1
36
7
1
37
7
1
38
4
1
39
4
1
40
4
1
41
4
1
42
3
1
43
3
1
44
3
1
45
3
1
46
8
1
47
8
1
48
8
1
49
8
1
50
1
1
51
1
1
52
1
1
53
1
1
54
7
1
55
7
1
56
7
1
57
7
1
58
12
1
59
12
1
60
12
1
61
12
1
62
9
1
63
9
1
64
9
1
65
9
1
66
9
1
67
9
1
68
9
1
69
9
1
70
8
1
71
8
1
72
8
1
73
8
1
74
6
1
75
6
1
76
6
1
77
6
1
78
9
1
79
9
1
80
9
1
81
9
1
82
12
1
83
12
1
84
12
1
85
12
1
86
8
1
87
8
1
88
8
1
89
8
1
90
9
1
91
9
1
92
9
1
93
9
1
94
15
1
95
15
1
96
15
1
97
15
1
98
9
1
99
9
End of preview. Expand in Data Studio

VizDoom (parquet)

GameNGen-style VizDoom agent gameplay with per-frame discrete actions, re-keyed from arnaudstiegler/vizdoom-50-episodes-skipframe-4 into the uniform per-frame parquet schema (one split per game; here doom).

~250k 320x240 RGB frames across 50 episodes. Each frame stores the discrete VizDoom action taken at that frame (the 18-way button set, ids [0, 17]).

Schema

  • image — the RGB game frame (Image, JPEG), 320x240
  • episode — episode id (windows never cross it)
  • frame_index — frame position within the episode
  • action — discrete VizDoom action id in [0, 17]

Usage

from datasets import load_dataset
doom = load_dataset("blanchon/doom_parquet", split="doom")
doom[0]["image"], doom[0]["action"]

Derived from arnaudstiegler/vizdoom-50-episodes-skipframe-4; redistributed for research.

Downloads last month
51