Update README.md
Browse files
README.md
CHANGED
|
@@ -86,19 +86,3 @@ sentence_embeddings = F.normalize(sentence_embeddings, p=2, dim=1)
|
|
| 86 |
print("Sentence embeddings:")
|
| 87 |
print(sentence_embeddings)
|
| 88 |
```
|
| 89 |
-
|
| 90 |
-
# Evaluation
|
| 91 |
-
|
| 92 |
-
The performance of the pre-trained model was evaluated using [ScandEval](https://github.com/ScandEval/ScandEval).
|
| 93 |
-
|
| 94 |
-
| Task | Dataset | Score (±SE) |
|
| 95 |
-
|:-------------------------|:-------------|:--------------------------------|
|
| 96 |
-
| sentiment-classification | angry-tweets | mcc = 36.14 (±1.07) |
|
| 97 |
-
| | | macro_f1 = 56.57 (±0.84) |
|
| 98 |
-
| named-entity-recognition | dane | micro_f1 = 55.56 (±1.69) |
|
| 99 |
-
| | | micro_f1_no_misc = 57.44 (±1.9) |
|
| 100 |
-
| linguistic-acceptability | scala-da | mcc = 12.4 (±3.07) |
|
| 101 |
-
| | | macro_f1 = 53.54 (±2.15) |
|
| 102 |
-
| question-answering | scandiqa-da | em = 17.87 (±1.15) |
|
| 103 |
-
| | | f1 = 27.84 (±1.37) |
|
| 104 |
-
| speed | speed | speed = 18.59 (±0.05) |
|
|
|
|
| 86 |
print("Sentence embeddings:")
|
| 87 |
print(sentence_embeddings)
|
| 88 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|