Update README.md
Browse files
README.md
CHANGED
@@ -19,4 +19,11 @@ The model was trained as cross-encoder classification model with the objective t
|
|
19 |
## Data
|
20 |
|
21 |
The data from SQuAD 2.0 was sentence-split. The question + the sentence containing the answer was a positive example.
|
22 |
-
The question + the remaining sentence from the same Wikipedia passege were treated as hard negative examples.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
19 |
## Data
|
20 |
|
21 |
The data from SQuAD 2.0 was sentence-split. The question + the sentence containing the answer was a positive example.
|
22 |
+
The question + the remaining sentence from the same Wikipedia passege were treated as hard negative examples.
|
23 |
+
|
24 |
+
The table balow reports the classification results on the validation set.
|
25 |
+
|
26 |
+
|
27 |
+
| | accuracy | F1 |
|
28 |
+
|----------------|----------|----------|
|
29 |
+
|ALBERT-xxlarge | 97.05 | 84.14 |
|