Update README.md
Browse files
README.md
CHANGED
|
@@ -26,15 +26,21 @@ It achieves the following results on the evaluation set:
|
|
| 26 |
|
| 27 |
## Model description
|
| 28 |
|
| 29 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
|
| 31 |
## Intended uses & limitations
|
| 32 |
-
|
| 33 |
-
More information needed
|
| 34 |
|
| 35 |
## Training and evaluation data
|
| 36 |
|
| 37 |
-
More information
|
| 38 |
|
| 39 |
## Training procedure
|
| 40 |
|
|
|
|
| 26 |
|
| 27 |
## Model description
|
| 28 |
|
| 29 |
+
The aim of this model is to provide gender-neutral terms for job titles in Polish in single sentences. The optimal outcome looks like this:
|
| 30 |
+
Jestem pracownikiem tej firmy.
|
| 31 |
+
turns into
|
| 32 |
+
Jestem osobą pracowniczą tej firmy.
|
| 33 |
+
Sentences not containing such terms are not expected to change at all, for example:
|
| 34 |
+
Mam uroczego kotka.
|
| 35 |
+
turns into
|
| 36 |
+
Mam uroczego kotka.
|
| 37 |
|
| 38 |
## Intended uses & limitations
|
| 39 |
+
The model has only been fine-tuned for single-sentence inputs, so other types of inputs can be unstable. So far, it underperforms for low-frequency items, morphosyntactically complex cases and feminine nouns.
|
|
|
|
| 40 |
|
| 41 |
## Training and evaluation data
|
| 42 |
|
| 43 |
+
The model's performance has been evaluated in terms of neutralization attempt precision and recall, as well as Levenshtein distance from gold-standard items. More information on the evaluation outcomes can be found [here](https://github.com/ArielUW/IMLLA-FinalProject/blob/main/README.md).
|
| 44 |
|
| 45 |
## Training procedure
|
| 46 |
|