Datasets:
First Update
#1
by malex2 - opened
README.md
CHANGED
|
@@ -6,4 +6,21 @@ language:
|
|
| 6 |
tags:
|
| 7 |
- Law
|
| 8 |
pretty_name: BodyCam-VQA Dataset
|
| 9 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
tags:
|
| 7 |
- Law
|
| 8 |
pretty_name: BodyCam-VQA Dataset
|
| 9 |
+
---
|
| 10 |
+
# Dataset Overview
|
| 11 |
+
BWC-VideoText-359 is a dataset fully generated from open-source videos found on the Civilian Office for Police Accountability website.
|
| 12 |
+
This dataset contains three hundred fifty-nine one minute videos and transcripts for each video.
|
| 13 |
+
|
| 14 |
+
# Stats
|
| 15 |
+
| Dataset Type | NumberOfItems |
|
| 16 |
+
|--------------------|---------------|
|
| 17 |
+
| eval_videos | 71 |
|
| 18 |
+
| eval_transcripts | 71 |
|
| 19 |
+
| train_videos | 288 |
|
| 20 |
+
| train_transcripts | 288 |
|
| 21 |
+
|
| 22 |
+
# Format
|
| 23 |
+
Videos are located in either train_videos.zip/eval_videos.zip named via:
|
| 24 |
+
Video[index].mp4
|
| 25 |
+
|
| 26 |
+
The corresponding transcripts are located in train_transcripts.zip/eval_transcripts.zip, they can be matched with the index.
|