updated README
Browse files
README.md
CHANGED
|
@@ -34,4 +34,25 @@ configs:
|
|
| 34 |
data_files:
|
| 35 |
- split: train
|
| 36 |
path: data/train-*
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
data_files:
|
| 35 |
- split: train
|
| 36 |
path: data/train-*
|
| 37 |
+
task_categories:
|
| 38 |
+
- sentence-similarity
|
| 39 |
+
language:
|
| 40 |
+
- en
|
| 41 |
+
pretty_name: BGE Retrieval Dataset (7 Datasets)
|
| 42 |
+
size_categories:
|
| 43 |
+
- 100K<n<1M
|
| 44 |
---
|
| 45 |
+
|
| 46 |
+
## BGE-Retrieval-Data (7 Datasets)
|
| 47 |
+
This filtered version of `nthakur/bge-full-data` includes only 7 datasets of ~678K training pairs.
|
| 48 |
+
The distribution of each subset is shown below:
|
| 49 |
+
|
| 50 |
+
| Dataset | Training Pairs | Pos | Hard Negs |
|
| 51 |
+
| -------: | -----------: | -----: | ------: |
|
| 52 |
+
| msmarco_passage | 485823 | 1.1 | 25.0 |
|
| 53 |
+
| hotpotqa | 84516 | 2.0 | 20.0 |
|
| 54 |
+
| nq | 58568 | 1.0 | 98.5 |
|
| 55 |
+
| fever | 29096 | 1.3 | 20.0 |
|
| 56 |
+
| scidocsrr | 12655 | 1.6 | 19.7 |
|
| 57 |
+
| fiqa | 5500 | 2.6 | 15.0 |
|
| 58 |
+
| arguana | 4065 | 1.0 | 13.6 |
|