Update ReadME
Browse files
README.md
CHANGED
|
@@ -1,6 +1,34 @@
|
|
| 1 |
---
|
| 2 |
-
license: mit
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
| 4 |
## **Dataset Structure Overview**
|
| 5 |
M-BEIR dataset comprises two main components: Query Data and Candidate Pool.
|
| 6 |
Each of these sections consists of structured entries in JSONL format (JSON Lines), meaning each line is a valid JSON object. Below is a detailed breakdown of the components and their respective fields:
|
|
|
|
| 1 |
---
|
| 2 |
+
license: "mit"
|
| 3 |
+
pretty_name: "M-BEIR"
|
| 4 |
+
task_categories:
|
| 5 |
+
- text retrieval
|
| 6 |
+
- text-to-image
|
| 7 |
+
- image-to-text
|
| 8 |
+
- visual question answering
|
| 9 |
+
language:
|
| 10 |
+
- "en"
|
| 11 |
+
configs:
|
| 12 |
+
- config_name: main_data
|
| 13 |
+
data_files:
|
| 14 |
+
- split: train
|
| 15 |
+
path: "train/*.jsonl"
|
| 16 |
+
- split: val
|
| 17 |
+
path: "val/*.jsonl"
|
| 18 |
+
- split: test
|
| 19 |
+
path: "test/*.jsonl"
|
| 20 |
+
- split: cand_pool
|
| 21 |
+
path: "cand_pool/*.jsonl"
|
| 22 |
+
- config_name: instructions
|
| 23 |
+
data_files:
|
| 24 |
+
- split: all
|
| 25 |
+
path: 'instructions/*.tsv'
|
| 26 |
+
- config_name: qrels
|
| 27 |
+
data_files:
|
| 28 |
+
- split: all
|
| 29 |
+
path: 'qrels/*.txt'
|
| 30 |
---
|
| 31 |
+
|
| 32 |
## **Dataset Structure Overview**
|
| 33 |
M-BEIR dataset comprises two main components: Query Data and Candidate Pool.
|
| 34 |
Each of these sections consists of structured entries in JSONL format (JSON Lines), meaning each line is a valid JSON object. Below is a detailed breakdown of the components and their respective fields:
|