NathanGavenski commited on
Commit
99f7955
·
verified ·
1 Parent(s): f930fb3

Update croissant.json

Browse files
Files changed (1) hide show
  1. croissant.json +182 -109
croissant.json CHANGED
@@ -1,126 +1,199 @@
1
  {
2
- "@context": "https://w3id.org/croissant/context/v1",
3
- "type": "Dataset",
4
- "name": "Labyrinth-v0 5x5 size",
5
- "description": "This dataset is for the gymnasium-based environment Labyrinth-v0. It contains observations, actions, rewards, episode_starts and info for each entry across 32 train, 32 validation, 32 test, and 32 random structure configurations.",
6
- "keywords": ["imitation_learning", "agents", "gymnasium"],
7
- "license": "https://opensource.org/licenses/MIT",
8
- "creator": [
9
- {
10
- "name": "Nathan Gavenski",
11
- "email": "nathangavenski@gmail.com"
12
- }
13
- ],
14
- "distribution": [
15
- {
16
- "name": "dataset.tar.gz",
17
- "encodingFormat": "application/gzip",
18
- "contentUrl": "https://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_5x5/resolve/main/dataset.tar.gz"
19
  },
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
20
  {
21
- "name": "images.tar.gz",
22
- "encodingFormat": "application/gzip",
23
- "contentUrl": "https://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_5x5/resolve/main/images.tar.gz"
 
 
 
 
24
  },
25
  {
26
- "name": "Labyrinth-v0_5x5.py",
27
- "encodingFormat": "text/x-python",
28
- "contentUrl": "https://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_5x5/resolve/main/Labyrinth-v0_5x5.py"
 
 
 
 
 
29
  },
30
  {
31
- "name": "README.md",
32
- "encodingFormat": "text/markdown",
33
- "contentUrl": "https://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_5x5/resolve/main/README.md"
 
 
 
 
 
34
  }
35
  ],
36
- "data": [
37
  {
38
- "name": "Train Split",
39
- "description": "Training data with 32 unique labyrinth structures.",
40
- "fileObject": {
41
- "contentUrl": "https://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_5x5/resolve/main/dataset.tar.gz",
42
- "encodingFormat": "application/jsonl"
43
- },
44
- "recordSet": {
45
- "record": {
46
- "type": "Record",
47
- "fields": [
48
- { "name": "obs", "dataType": "string", "description": "Image path for observation." },
49
- { "name": "actions", "dataType": "integer", "description": "Action taken." },
50
- { "name": "rewards", "dataType": "float", "description": "Reward value." },
51
- { "name": "episode_starts", "dataType": "boolean", "description": "Is episode start." },
52
- { "name": "info", "dataType": "string", "description": "Environment metadata." }
53
- ]
54
  }
55
- }
 
 
 
 
 
 
56
  },
57
  {
58
- "name": "Validation Split",
59
- "description": "Evaluation data on unseen labyrinths.",
60
- "fileObject": {
61
- "contentUrl": "https://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_5x5/resolve/main/dataset.tar.gz",
62
- "encodingFormat": "application/jsonl"
63
- },
64
- "recordSet": {
65
- "record": {
66
- "type": "Record",
67
- "fields": [
68
- { "name": "obs", "dataType": "string", "description": "Image path for observation." },
69
- { "name": "actions", "dataType": "integer", "description": "Action taken." },
70
- { "name": "rewards", "dataType": "float", "description": "Reward value." },
71
- { "name": "episode_starts", "dataType": "boolean", "description": "Is episode start." },
72
- { "name": "info", "dataType": "string", "description": "Environment metadata." }
73
- ]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
74
  }
75
- }
76
- },
77
- {
78
- "name": "Test Split",
79
- "description": "Held-out data for final performance reporting.",
80
- "fileObject": {
81
- "contentUrl": "https://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_5x5/resolve/main/dataset.tar.gz",
82
- "encodingFormat": "application/jsonl"
83
- },
84
- "recordSet": {
85
- "record": {
86
- "type": "Record",
87
- "fields": [
88
- { "name": "obs", "dataType": "string", "description": "Image path for observation." },
89
- { "name": "actions", "dataType": "integer", "description": "Action taken." },
90
- { "name": "rewards", "dataType": "float", "description": "Reward value." },
91
- { "name": "episode_starts", "dataType": "boolean", "description": "Is episode start." },
92
- { "name": "info", "dataType": "string", "description": "Environment metadata." }
93
- ]
94
- }
95
- }
96
- },
97
- {
98
- "name": "Random Policy Split",
99
- "description": "Randomly collected data for methods that require random transitions (such as BCO, ABCO, and IUPE).",
100
- "fileObject": {
101
- "contentUrl": "hhttps://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_5x5/resolve/main/dataset.tar.gz",
102
- "encodingFormat": "application/jsonl"
103
- },
104
- "recordSet": {
105
- "record": {
106
- "type": "Record",
107
- "fields": [
108
- { "name": "obs", "dataType": "string", "description": "Image path for observation." },
109
- { "name": "actions", "dataType": "integer", "description": "Action taken." },
110
- { "name": "rewards", "dataType": "float", "description": "Reward value." },
111
- { "name": "episode_starts", "dataType": "boolean", "description": "Is episode start." },
112
- { "name": "info", "dataType": "string", "description": "Environment metadata." }
113
- ]
114
- }
115
- }
116
- },
117
- {
118
- "name": "Observation Images",
119
- "description": "All image frames grouped by split (train/, test/, eval/, random/).",
120
- "fileObject": {
121
- "contentUrl": "https://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_5x5/resolve/main/images.tar.gz",
122
- "encodingFormat": "image/jpeg"
123
- }
124
  }
125
  ]
126
  }
 
1
  {
2
+ "@context": {
3
+ "@language": "en",
4
+ "@vocab": "https://schema.org/",
5
+ "citeAs": "Nathan Gavenski. Labyrinth-v0 4x4 size. https://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_4x4",
6
+ "column": "cr:column",
7
+ "conformsTo": "dct:conformsTo",
8
+ "cr": "http://mlcommons.org/croissant/",
9
+ "data": {
10
+ "@id": "cr:data",
11
+ "@type": "@json"
12
+ },
13
+ "dataBiases": "cr:dataBiases",
14
+ "dataCollection": "cr:dataCollection",
15
+ "dataType": {
16
+ "@id": "cr:dataType",
17
+ "@type": "@vocab"
 
18
  },
19
+ "dct": "http://purl.org/dc/terms/",
20
+ "extract": "cr:extract",
21
+ "field": "cr:field",
22
+ "fileProperty": "cr:fileProperty",
23
+ "fileObject": "cr:fileObject",
24
+ "fileSet": "cr:fileSet",
25
+ "format": "cr:format",
26
+ "includes": "cr:includes",
27
+ "isLiveDataset": "cr:isLiveDataset",
28
+ "jsonPath": "cr:jsonPath",
29
+ "key": "cr:key",
30
+ "md5": "cr:md5",
31
+ "parentField": "cr:parentField",
32
+ "path": "cr:path",
33
+ "personalSensitiveInformation": "cr:personalSensitiveInformation",
34
+ "recordSet": "cr:recordSet",
35
+ "references": "cr:references",
36
+ "regex": "cr:regex",
37
+ "repeated": "cr:repeated",
38
+ "replace": "cr:replace",
39
+ "sc": "https://schema.org/",
40
+ "separator": "cr:separator",
41
+ "source": "cr:source",
42
+ "subField": "cr:subField",
43
+ "transform": "cr:transform"
44
+ },
45
+ "@type": "sc:Dataset",
46
+ "conformsTo": "http://mlcommons.org/croissant/1.0",
47
+ "name": "Labyrinth-v0 4x4 size",
48
+ "description": "This dataset is for the Gymnasium-based environment Labyrinth-v0. It contains observations, actions, rewards, episode_starts and info across multiple structural splits (train, test, eval, random).",
49
+ "alternateName": ["NathanGavenski/Labyrinth-v0_4x4", "Labyrinth-v0"],
50
+ "url": "https://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_4x4",
51
+ "license": "https://choosealicense.com/licenses/mit/",
52
+ "creator": {
53
+ "@type": "Person",
54
+ "name": "Nathan Gavenski",
55
+ "url": "https://huggingface.co/NathanGavenski"
56
+ },
57
+ "keywords": ["Gymnasium", "Imitation Learning", "Expert Trajectory", "JSONL", "Image-based Agent", "4x4 Labyrinth"],
58
+ "distribution": [
59
  {
60
+ "@type": "cr:FileObject",
61
+ "@id": "repo",
62
+ "name": "repo",
63
+ "description": "The Hugging Face git repository for Labyrinth-v0_4x4.",
64
+ "contentUrl": "https://huggingface.co/datasets/NathanGavenski/Labyrinth-v0_4x4/tree/main",
65
+ "encodingFormat": "git+https",
66
+ "sha256": "main"
67
  },
68
  {
69
+ "@type": "cr:FileSet",
70
+ "@id": "dataset-files",
71
+ "name": "dataset-files",
72
+ "description": "The JSONL split files for Labyrinth-v0 4x4 dataset.",
73
+ "containedIn": { "@id": "repo" },
74
+ "encodingFormat": "application/jsonl",
75
+ "includes": "dataset/*_shortest_route.jsonl",
76
+ "sha256": "04dd930713199dd3e2e2239e981de3e15b2ea8ac715bb5349e9058991f3a9bca"
77
  },
78
  {
79
+ "@type": "cr:FileSet",
80
+ "@id": "images-folder",
81
+ "name": "images-folder",
82
+ "description": "Image files grouped by split.",
83
+ "containedIn": { "@id": "repo" },
84
+ "encodingFormat": "image/jpeg",
85
+ "includes": "images/*/*.jpg",
86
+ "sha256": "1de30325f521bcdaa19254a5b2339eaba472d444cb47d49426051fc78aa39e9f"
87
  }
88
  ],
89
+ "recordSet": [
90
  {
91
+ "@type": "cr:RecordSet",
92
+ "@id": "default_splits",
93
+ "name": "Split list",
94
+ "description": "Split definitions for Labyrinth dataset.",
95
+ "dataType": "cr:Split",
96
+ "key": { "@id": "default_splits/split_name" },
97
+ "field": [
98
+ {
99
+ "@type": "cr:Field",
100
+ "@id": "default_splits/split_name",
101
+ "name": "split_name",
102
+ "description": "Split identifier.",
103
+ "dataType": "sc:Text"
 
 
 
104
  }
105
+ ],
106
+ "data": [
107
+ { "default_splits/split_name": "train" },
108
+ { "default_splits/split_name": "eval" },
109
+ { "default_splits/split_name": "test" },
110
+ { "default_splits/split_name": "random" }
111
+ ]
112
  },
113
  {
114
+ "@type": "cr:RecordSet",
115
+ "@id": "default",
116
+ "name": "Labyrinth-v0 4x4 Interaction Data",
117
+ "description": "Main dataset fields for each step in the environment.",
118
+ "field": [
119
+ {
120
+ "@type": "cr:Field",
121
+ "@id": "default/split",
122
+ "name": "default/split",
123
+ "description": "Split to which the entry belongs.",
124
+ "dataType": "sc:Text",
125
+ "source": {
126
+ "fileSet": { "@id": "dataset-files" },
127
+ "extract": { "fileProperty": "fullpath" },
128
+ "transform": {
129
+ "regex": "dataset/(train|test|eval|random)_shortest_route.jsonl"
130
+ }
131
+ },
132
+ "references": {
133
+ "field": { "@id": "default_splits/split_name" }
134
+ }
135
+ },
136
+ {
137
+ "@type": "cr:Field",
138
+ "@id": "default/obs",
139
+ "name": "default/obs",
140
+ "description": "Path to the image representing the observation.",
141
+ "dataType": "sc:Text",
142
+ "source": {
143
+ "fileSet": { "@id": "dataset-files" },
144
+ "extract": { "column": "obs" }
145
+ },
146
+ "repeated": true
147
+ },
148
+ {
149
+ "@type": "cr:Field",
150
+ "@id": "default/actions",
151
+ "name": "default/actions",
152
+ "description": "Action taken at this step.",
153
+ "dataType": "sc:Integer",
154
+ "source": {
155
+ "fileSet": { "@id": "dataset-files" },
156
+ "extract": { "column": "actions" }
157
+ },
158
+ "repeated": true
159
+ },
160
+ {
161
+ "@type": "cr:Field",
162
+ "@id": "default/rewards",
163
+ "name": "default/rewards",
164
+ "description": "Reward received.",
165
+ "dataType": "sc:Float",
166
+ "source": {
167
+ "fileSet": { "@id": "dataset-files" },
168
+ "extract": { "column": "rewards" }
169
+ },
170
+ "repeated": true
171
+ },
172
+ {
173
+ "@type": "cr:Field",
174
+ "@id": "default/episode_starts",
175
+ "name": "default/episode_starts",
176
+ "description": "Whether this is the first step of the episode.",
177
+ "dataType": "sc:Boolean",
178
+ "source": {
179
+ "fileSet": { "@id": "dataset-files" },
180
+ "extract": { "column": "episode_starts" }
181
+ },
182
+ "repeated": true
183
+ },
184
+ {
185
+ "@type": "cr:Field",
186
+ "@id": "default/info",
187
+ "name": "default/info",
188
+ "description": "Information needed to reconstruct the labyrinth state.",
189
+ "dataType": "sc:Text",
190
+ "source": {
191
+ "fileSet": { "@id": "dataset-files" },
192
+ "extract": { "column": "info" }
193
+ },
194
+ "repeated": true
195
  }
196
+ ]
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
197
  }
198
  ]
199
  }