JinPLU commited on
Commit
3836ce3
·
verified ·
1 Parent(s): 9ab987f

Upload folder using huggingface_hub

Browse files
README.md ADDED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ pretty_name: "WRBench Human Annotations"
3
+ license: apache-2.0
4
+ language:
5
+ - en
6
+ tags:
7
+ - video
8
+ - benchmark
9
+ - human-evaluation
10
+ - wrbench
11
+ - arxiv:2606.20545
12
+ task_categories:
13
+ - image-to-video
14
+ size_categories:
15
+ - n<1K
16
+ configs:
17
+ - config_name: pairs
18
+ data_files:
19
+ - split: default
20
+ path: data/pairs.parquet
21
+ - config_name: annotation_to_video
22
+ data_files:
23
+ - split: default
24
+ path: data/annotation_to_video.parquet
25
+ - config_name: exclusions
26
+ data_files:
27
+ - split: default
28
+ path: data/exclusions.parquet
29
+ ---
30
+
31
+ # WRBench Human Annotations
32
+
33
+ Scrubbed human pairwise-comparison metadata mapped to public `video_asset_id`
34
+ values. Release boundary: **1,156 clean pairs** with **2,547
35
+ annotator verdict records**. **1,485 maintenance-queue pairs are excluded**
36
+ (see `exclusions` config).
37
+
38
+ ## Load
39
+
40
+ ```python
41
+ from datasets import load_dataset
42
+
43
+ pairs = load_dataset("WRBench/wrbench-human-annotations", "pairs", split="default")
44
+ ```
45
+
46
+ ## Privacy
47
+
48
+ Public fields only: `pair_id`, anonymized `verdict_annotators`, timestamps,
49
+ `video_asset_id`, and release-safe endpoint metadata. No internal paths or PII.
50
+
51
+ ## Links
52
+
53
+ - Videos: https://huggingface.co/datasets/WRBench/wrbench-videos
data/annotation_to_video.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a1652cf4a8592e9cd8db261e526e3405b0e35c1980dd0bae592f5a8e1c09e12
3
+ size 82613
data/exclusions.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e05158076b61262032848396fa138da5c10d340627d9ef137688eb48e0e34b56
3
+ size 13117
data/pairs.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a8dfef1fa648741e4da47c319ad5c85f13cd1bf7806a50147025b834b042b536
3
+ size 163592
datasheet.md ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # WRBench Human Annotations Datasheet
2
+
3
+ ## Motivation
4
+
5
+ Human pairwise comparisons supporting WRBench diagnostic claims.
6
+
7
+ ## Composition
8
+
9
+ - Clean pairs released: 1156
10
+ - Verdict records (sum of n_verdict_annotators): 2547
11
+ - Excluded maintenance pairs: documented in exclusions config
12
+
13
+ ## Collection
14
+
15
+ Annotators compared model outputs on controlled pairwise tasks. Identities are
16
+ pseudonymized (`p1`, `p2`, ...).
17
+
18
+ ## Recommended uses
19
+
20
+ Benchmark analysis, calibration of automated metrics, qualitative error studies.
21
+
22
+ ## Out of scope
23
+
24
+ Training annotator models; deanonymization attempts.
original/annotation_to_video.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
original/exclusions.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
original/pairs.jsonl ADDED
The diff for this file is too large to render. See raw diff